Ernst Schering Research Foundation Workshop 48 From Morphological Imaging To Molecular Targeting

Ernst Schering Research Foundation Workshop 48
From Morphological Imaging to Molecular Targeting

Springer-Verlag Berlin Heidelberg GmbH
Ernst Schering Research Foundation
Workshop 48
From Morphological Imaging

to Molecular Targeting
Implications to Preclinical Development
M. Schwaiger, L. Dinkelborg, H. Schweinfurth

Editors
With 20 Figures
Springer
Series Editors: G. Stock and M. Lessl
ISSN 0947-6075
ISBN 978-3-662-07312-4
Library of Congress Cataloging-in-Publication Data
From morphological imaging to molecular targeting: implications for preclinical
developmentIM. Schwaiger, L. Dinkelborg, and H. Schweinfurth.
p.; em. - (Ernst Schering Research Foundation workshop, ISSN 0947-6075; 48)
Includes bibliographical references and index.
ISBN 978-3-662-07312-4 ISBN 978-3-662-07310-0 (eBook)
DOI 10.1007/978-3-662-07310-0
1. Radiopharmaceuticals - Congresses. 2. Drug targeting - Congresses. 3. Ligand binding
(Biochemistry) - Congresses. 4. Drugs - Design - Congresses. 1. Schwaiger, Markus. II.
Dinkelborg, L. (Ludger), 1962- III. Schweinfurth, H. (Hennann), 1950- IV. Series.
[DNLM: I. Drug Evaluation, Preclinical - Congresses. 2. Radiopharmaceuticals -
Congresses. 3. Clinical Trials - Congresses. 4. Drug Delivery Systems - methods -
Congresses. 5. Drug Design - Congresses. QV 771 M871 2004]
This work is subject to copyright. All rights are reserved, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data
banks. Duplication of this publication or parts thereof is permitted only under the provisions
of the Gennan Copyright Law of September 9, 1965, in its current version, and permission
for use must always be obtained from Springer-Verlag Berlin Heidelberg GmbH. Violations are
liable for prosecution under the Gennan Copyright Law.
springeronline.com
© Springer-Verlag Berlin Heidelberg 2004
Originally published by Springer-Verlag Berlin Heidelberg New York in 2004
Softcover reprint of the hardcover 1st edition 2004
The use of general descriptive names, registered names, trademarks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use. Product
liability: The publishers cannot guarantee the accuracy of any information about dosage and
application contained in this book. In every individual case the user must check such infor-
mation by consulting the relevant literature.
Cover design: design & production GmbH, Heidelberg
Typesetting: K + V Fotosatz GmbH, Beerfelden
2113l30/AG-5 4 3 2 1 0 - Printed on acid-free paper
Preface
Because of the current progress in molecular medicine (genomics,

proteomics), a plethora of new and often human-specific targets are
being identified. These targets often play a significant role in the
pathogenesis of diseases, and identifying them offers the potential
for early diagnosis and intervention. An early in vivo validation of
specific ligands binding to these targets in humans is needed to as-
sess their potential for targeted imaging and radiotherapy. Further-
VI Preface
more, such validation studies may allow for a better understanding

of the molecular processes underlying phannacologic activity and
therefore for a more successful development of phannaceuticals in
general. The purpose of the Ernst Schering Research Foundation
(ESRF) Workshop 48 was to provide a forum for an open exchange
on the state of the art in the early development of such radiophanna-
ceuticals. Experts from academia, industry, and regulatory authori-
ties were invited to give presentations on aspects covering the identi-
fication of targets, preclinical studies on the safety of ligands, as
well as their validation in human clinical trials. It was our intention
to cover both the opportunities and the challenges that scientists in
this field are facing.
Radiopharmaceuticals are uniquely suitable for the above-men-
tioned target validation studies. Because of the high detection sensi-
tivity of the applied imaging technology (gamma camera, PET scan-
ner), the amount of administered compound in clinical trials is
usually very small and far below the threshold for any expected
phannacological effects. Still, there are a number of caveats that in-
vestigators should be aware. of when they intend to move towards
target validation in humans. At the ESRF Workshop, clinical experts
presented their experience in this rapidly developing field and their
approaches with regard to the preclinical characterization.
Above all, one has to make sure that there are no unacceptable
safety concerns with regard to the exposure of the participants in
these clinical trials. To achieve this goal, the preclinical program de-
signed for such a new drug substance will have to take into account
the scientific state of the art as well as the requirements laid down
by regulatory authorities. If this is not done properly, such a study in
human subjects could be unethical and may be even illegal.
It is very obvious that safety studies should be selected with the
particular nature of the radiophannaceutical agent in mind. Thus in
the first instance, for an antibody an in vitro immunohistochemical
study for cross-reactivity should be conducted in a panel of nonnal
human tissues. It is suggested to include animal tissues as well. The
results of such experiments should either provide a rationale for the
selection of a suitable animal species for in vivo preclinical studies
or support the argument that appropriate animal models (i.e., for de-
tection of consequences of interaction with the intended target) are
Preface VII
not available for safety testing. Even in the latter case, an in vivo
study may still be appropriate to demonstrate that no unexpected
side effects based on the properties of the new ligand have to be
reckoned with, for example, as a result of impurities. Furthermore,
the tissue distribution of a new radiopharmaceutical should be stud-
ied in experimental animals to provide a basis for the assessment of
exposure of target and nontarget tissues.
This leads us to the area of radiation safety assessment. Even if
the dose of radioactivity is generally very small for a radiodiagnos-
tic agent, the internal radiation dose absorbed by organs should be
estimated according to an accepted method such as the Medical In-
ternal Radiation Dose (MIRD) system. One chapter in this volume is
dedicated to the background of the MIRD methodology and its ap-
plication in the extrapolation from experimental animals to human
beings.
To complicate the situation even more, the safety requirements
for such clinical trials are dealt with in a variety of regulatory guide-
lines, some of them being of international character (i.e., those
agreed upon within the International Conference on Harmonization,
ICH) and some being of regional importance (depending on where
the clinical trial will take place). The general requirements for pre-
clinical safety tests prior to clinical trials have been laid down in the
ICH Guideline M3 ("Maintenance of the ICH Guideline on Noncli-
nical Safety Studies for the Conduct of Human Clinical Trials for
Pharmaceuticals", 9 November 2000). If a ligand is derived from
biotechnology, the more specific guidance of the ICH Guideline S6
("Preclinical Safety Evaluation of Biotechnology-Derived Pharma-
ceuticals", July 1997) will also have to be considered. The FDA has
recently published a new draft guidance for nonbiological and bio-
logical imaging agents, which opens the avenue to initiate phase I
clinical trials without having to conduct comprehensive repeated-
dose toxicity studies. However, while this will keep the consumption
of test substance relatively low, there may still be concern about the
resources needed for the required single-dose toxicity studies with
extended examination in two species. A similar approach to that of
the FDA is described for the EU in a Committee for Proprietary
Medicinal Products (CPMP) position paper on studies to support
early clinical trials with a single low dose. Although the reader may
VIII Preface
find that this is exactly what should apply to proof-of-concept stud-

ies with radiopharmaceuticals, one might be disappointed by the re-
strictions under which this guideline is applicable. Thus, a ligand
molecule should not be a "biotech compound" (since, otherwise, one
would be referred to the above-mentioned Guideline ICH-S6) and
the intended human dose should not exceed a maximum of 100 J.lg.
Therefore, it seems that further discussions with the regulatory
authorities will be needed to define the preclinical safety require-
ments which are appropriate for radiopharmaceutical ligands. Read-
ers interested in the regulatory environment for radiopharmaceuticals
will learn much more about these issues in the sections addressing
regulatory aspects for the EU, the USA, and Japan.
At the ESRF Workshop there was open communication on the
subjects mentioned above among the experts from the preclinical,
clinical, and regulatory fields. The editors hope that the publication
of this volume will help to foster further discussions that are neces-
sary to promote the promising field of radiopharmaceutical develop-
ment.
Markus Schwaiger
Ludger Dinkelborg
Hermann Schweinfurth
Contents
Molecular Imaging: Dream or Reality?

M. Schwaiger, W Weber ...................... .
2 Target Discovery and Validation

R. Solari . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3 Noninvasive Imaging in Drug Discovery and Development

M. Rudin, P. Allegrini, N. Beckmann, H.-Ulrich Gremlich,
R. Kneuer, D. Laurent, M. Rausch, M. Stoeckli ....... 47
4 Internal Dose Assessment -

Extrapolation from Animal Species to the Human
M. G. Stabin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5 PET for Drug Development

C. Halldin, B. Gulyas, L. Farde 95
6 Future Directions in Molecular Imaging

V. Haberkorn .......... . . . . . . . . . . . . . . . . . . . .. 111
7 Regulatory Implications - The EU Perspective

A. Neil. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 135
8 Nonclinical Development of Radiopharmaceuticals:

Regulatory Considerations for the United States Food
and Drug Administration
S. Wilson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
x Contents
9 The Japanese Perspective on Radiopharmaceuticals

H. Mayahara .............................. . 167
Previous Volumes Published in This Series ............. 193

List of Editors and Contributors
Editors
Schwaiger, M.
Technical University Munich, Ismaninger Str. 221
Klinikum rechts der Isar, 81675 Munich, Germany
e-mail: m.schwaiger@lrz.tu-muenchen.de
Dinkelborg, L.
Schering AG, Radiopharmaceuticals Research, Miillerstr. 178,
13342 Berlin, Germany
e-mail: ludger.dinkelborg@schering.de
Schweinfurth, H.
Schering AG, Experimental Toxicology, 13342 Berlin, Germany
e-mail: hermann.schweinfurth@schering.de
Contributors
Allegrini, P.
Novartis Institute for Biomedical Research, Analytical and Imaging
Sciences Unit,WKL-125.516, 4002 Basel, Switzerland
e-mail: Peter.allegrini@pharma.novartis.com
Beckmann, N.
Sciences Unit, WKL-125.5.l6, 4002 Basel, Switzerland
XII List of Editors and Contributors
Farde, L.
Karolinska Institute, Department of Clinical Neuroscience,
Psychiatry Section, Karolinska Hospital, 17176 Stockholm, Sweden
e-mail: lars.fared@ks.se
Gremlich, H.-u.
Sciences Unit, WSJ-386.1O.54, 4002 Basel, Switzerland
e-mail: hansulrich.gremlich@pharma.novartis.com
Gulyas, B.
e-mail: balazs.gulyas@neuro.ki.se
Haberkorn, U.
Universitat Heidelberg DKFZ, 1m Neuenheimer Feld 400,
69120 Heidelberg, Germany
e-mail: uweJIaberkorn@med.uni-heidelberg.de
Halldin, C.
e-mail: christer.halldin@ks.se
Kneuer, R.
Sciences Unit, WSJ-507.5.04, 4002 Basel, Switzerland
Laurent, D.
Novartis Institute for Biomedical Research, 437-l319,
Novartis Pharmaceutical Corporation, One Health Plaza,
East Hanover, NJ 07936-1080, USA
e-mail: Didier.laurent@pharma.novartis.com
Mayahara, H.
International Clinical Research Organization for Medicine
(InCROM) Institute, 4-12-11, Kaduga, Suita Osaka 565-0853, Japan
e-mail: mayahara@incrom.co.jp
List of Editors and Contributors XIII
Neil, A.
Division of Clinical Trials, Medicinal Products Agency,
P.O. Box 26, 75103 Uppsala, Sweden
e-mail: anders.neil@mpa.se
Rausch, M.
Sciences Unit, 4002 Basel, Switzerland
Rudin, M.
e-mail: markus-l.rudin@pharma.novartis.com
Solari, R.
Medical Research Council Technology,
20 Park Crescent, London WIB lAL, UK
e-mail: roberto.solari@tech.rnrc.ac.uk
Stab in, M.G.

Vanderbilt University Medical Center Department of Radiology
and Radiological Sciences, 1161 21st Avenue South,
Nashville TN 372323-2675, USA
e-mail: Michael.g.stabin@vanderbilt.edu
Stoeckli, M.
e-mail: markus.stoeckli@pharma.novartis.com
Weber, W.
Technical University Munich, Ismaninger Str. 27,
81675 Munich, Germany
Wilson, S.
Milestone Biomedical Associates, 15 Worman's Mill Court, Suite 1,
Frederick, MD 21701, USA
e-mail: susanwilson@criver.com
1 Molecular Imaging: Dream or Reality?
M. Schwaiger, W. Weber
One has to try the impossible to achieve the possible.

(Hermann Hesse)
1.1 Imaging Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2 Imaging Targets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Visualization of Target Proteins . . . . . . . . . . . . . . . . . . . . . 7
1.4 Assessment of Protein Function . . . . . . . . . . . . . . . . . . . . 9
1.5 Biological Imaging as a Clinical Approach . . . . . . . . . . . . . 10
1.6 Monitoring of Therapy Response. . . . . . . . . . . . . . . . . . . . 12
1.7 Conclusion.................................. 13
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Imaging has witnessed considerable advances during the last

20 years. Morphological imaging with computed tomography (CT)
and magnetic resonance (MR) has benefited from the improvement
of spatial and temporal resolution, which revitalized especially the
role of CT in diagnostic imaging. Multislice CT (MSCT) imaging
has become an important tool for staging and restaging of oncologi-
cal diseases as well as a new diagnostic modality to detect calcifica-
tions of the coronary arteries. At the same time, functional imaging
has expanded its role in the clinical assessment of severity and ex-
tent of disease processes with documented prognostic value. MR
imaging as well as tracer techniques using single photon emission
2 M. Schwaiger, W. Weber
computed tomography (SPECT) and positron emission tomography

(PET) have added an important new dimension to the functional
characterization of pathophysiological processes by providing re-
gional information on perfusion, metabolism, and cell integrity.
More recently, with the increasing knowledge in molecular biology,
new imaging targets have been identified. Specific receptor families
as well as cell surface proteins have been proposed as targets for
various imaging approaches using radiolabeled peptides, antibodies,
or MR contrast agents. The specific noninvasive visualization of pro-
tein expression has become possible for diagnosis and therapy guid-
ance. Using transgenic approaches, specific proteins can be ex-
pressed, providing reporter gene imaging for the visualization of
gene expression. In this lecture I shall address the characteristics of
established and emerging imaging technologies as well as discuss
possible applications of molecular imaging (Phelps 2000).
1.1 Imaging Technologies
Many new and established imaging modalities compete for clinical

and laboratory applications (Fig. 1). This healthy rivalry supports the
continuing methodological development of each of the imaging tech-
niques. Most recently, the combination of morphological as well as
functional imaging has been proposed. PET/CT and SPECT/CT in-
strumentation have been commercially introduced and rapidly ac-
cepted in the oncological imaging community. Combining metabolic
imaging with PET and high-resolution MSCT imaging allows unique
colocalization of metabolic activity, resulting in improved staging
and restaging of patients with oncological diseases. Therefore, the
future of multi modality imaging opens new directions in clinical and
research applications, which will have major impact on in vivo char-
acterization of biological systems.
Imaging technologies differ by important specific performance
characteristics. As mentioned above, MR imaging and CT excel by
high spatial resolution, while the tracer techniques provide high sen-
sitivity, which allows the detection of picomolar concentrations of
radiolabeled compounds in the body. Optical imaging is a sensitive
technique, coupled with high, but depth-dependent resolution (Con-
Molecular Imaging: Dream or Reality? 3
Anatomy X-Ray
Computer Tomography (CT)
Contrast- Angiography
kinetics Ulrasound ISPECTI PET
Magnetic resonance Imaging
(MRI)
Metabolism
Receptors Tracer Technique
<.,..SPECT,.,..PET)
Signal Transduction Optical

Imaging
Cell Trafficking
Biology
Fig. 1. Currently available imaging techniques and biological processes un-
der investigation. There is continuous development from structural toward
the biological interrogation of tissue by imaging techniques
tag et a1. 1998). The penetration of light in tissue is a complex diffu-

sion process and, therefore, only signals very close to the surface or
requires an invasive optical imaging device for accurate assessments.
A possible application in the near future will be optical imaging
probes, which can be advanced into, e.g., coronaries (optical coher-
ence tomography, OCT) or the gastrointestinal tract for local detec-
tion of fluorescent signals.
Recently, the term "molecular imaging" has been introduced to
emphasize new targets of functional imaging. The goal of "molecu-
lar", "biological", "physiological", or "medical" imaging is to visua-
lize biological processes in vivo and to relate them to diagnosis,
prognosis, and therapy in a given patient. These clinical applications
contrast with the experimental strategies, in which imaging will play
an increasing role to quantitatively assess biological processes in an-
imal models. This will be helpful for the better understanding of
pathophysiology as well as characterizing functionally the phenotype
of transgenic mice. Additionally, imaging will gain increasing im-
portance in the preclinical evaluation of new therapeutic approaches.
The spectrum of imaging techniques available today provides both
clinically and experimentally attractive information of structure, phy-

siology, biochemistry, and molecular biology.
What is the future of molecular imaging? We will most likely ob-
serve a branching of exciting developments into clinical and experi-
mental applications. Clinical imaging will most likely be dominated
by tracer techniques, which are able to produce a large number of
validated biological imaging signals in vivo. The main focus of this
application will be "translational research" (Jacobs et al. 2002), in
vivo phenotyping (de Vries and Vaalburg 2002), and drug evalua-
tion. It is expected that the effect of pharmacological therapy can be
assessed specifically based on functional parameters using tracer
techniques. Based on available research results, it is expected that
molecular imaging, combined with high-resolution structural imag-
ing will be an important part of diagnosis, therapy guidance, and
therapy monitoring in the near future. On the other hand, the labora-
tory research will primarily rely on dedicated microimaging devices
including micro CT, micro PET, and MR as well as optical imaging
Chatziioannou et al. 1999; Ziegler et al. 2001; Knoess et al. 2003,
because it offers advantages in terms of required infrastructure, ex-
pense, and availability of nonradioactive imaging probes. Interdisci-
plinary imaging laboratories will be established providing central
service units for image acquisition and processing to support basic
science as well as preclinical drug development.
1.2 Imaging Targets
What are the currently employed strategies of molecular imaging

(Fig. 2)? Among many exciting new techniques, I will address repor-
ter gene imaging (Bogdanov and Weissleder 1998), visualization of
target proteins, and assessment of protein function. In addition, clini-
cal imaging using biological signals for monitoring therapy response
will be briefly discussed.
Reporter gene imaging has been pioneered by Dr. Blasberg's
group at Sloan Kettering Cancer Center in New York (Tjuvajev et
al. 1995), which introduced the reporter gene approach with the
HSVlIthymidine kinase (tk) gene in 1995. Herpes viral tk can be de-
tected in mammal cells using a radiolabeled thymidine analog repor-
• Reporter Gene Imaging

• Visualization of target proteins
• Assessment of protein function
• Clinical imaging using biological signals
• Monitoring of therapy response
Fig. 2. Possible targets for imaging from the gene to biochemical or physio-
logical function
ter substrate (FIAU), which is phosphorylated by tk and retained in

tissue. With this tracer approach, the protein expression can be quan-
titated noninvasively as shown by several studies comparing imaging
information with in vitro assessment of protein concentration. Tjuva-
jev et aI. (1999) have demonstrated that there is a close correlation
between gene expression and the scintigraphic signal using PET
imaging in combination with iodine-124 FIAU. Similar data were
published by the UCLA group of Gambhir et aI., using a mutated tk
glue and a radiolabeled analog of the ganciclovir (Gambhir et aI.
1998; Yaghoubi et aI. 2001). The aim of this innovative imaging
approach is to combine a reporter gene with a therapeutic gene. By
measuring the intensity of tracer uptake, the transfection efficiency
can be noninvasively monitored and related to regional therapeutic
effects. Aside from tracer approaches, fluorescent probes are unique-
ly suited to follow gene expression in cells and small animal
preparations (Lansford et aI. 2001; Blasberg and Gelovani 2002). In
addition, MRI and MRS have been proposed as monitoring devices
in transgenic models (Stegman et aI. 1999; Bell and Taylor-Robin-
son 2000; Weissleder and Mahmood 2001).
More recently, reporter gene imaging has been used to visualize
activation of specific genes. Doubrovin et aI. (2001) reported the
imaging of the activated tumor suppressor gene p53. A p53 enhancer
was used controlling tk gene expression, which resulted in specific
tissue gene expression in the presence of p53. Imaging prior to and
during chemotherapy, which induced the expression of p53, resulted
in the visualization of regional expression of tk protein in the trans-
fected tumor.
Aside from the tk reporter gene, other reporter gene approaches
have been introduced. The expression of sodium-iodide transporter
(NIS) proposed by Dai et al. represents a promising tool for radio-

therapy with 1-131 (Dai et al. 1996). Imaging studies in various ani-
mal models demonstrated that the expression of this transporter can
be monitored with radioactive iodine, but the half life of 1-123 with-
in the transfected tissue is relatively short. Future studies have to
show if the reporter gene approach using NIS is of experimental or
clinical utility (Haberkorn and Altmann 2001).
Most recently, Weber et al. from our group transfected cells with
the norepinephrine transporter (NET) as reporter gene. The norepi-
nephrine transporter is a member of amine transporters being ex-
pressed primarily in sympathetic nerve terminals. NET-transfected
tumor cells can be visualized using a false neurotransmitter such as
1-123 MIBG. In contrast to NIS, the cells retain the catecholamine
analog over prolonged periods. Using this approach, the question
arises if the biological contrast between transfected cells and normal
tissue, which may be innervated by sympathetic nerves, is high
enough to produce a useful diagnostic or therapeutic approach (We-
ber et al. 2003).
Recently, Bengel et al. from our group introduced the HSV-tk-
gene PET imaging approach to the heart and demonstrated for the
first time that in vivo transfection in the rat as well as in the pig
model can be successfully used to monitor the expression of the re-
porter gene in using 1-124 FIAU (Bengel et al. 2000). The area of
transfection can be localized with high spatial resolution and corre-
lated with tissue function imaging such as blood flow and metabo-
lism. In contrast to tumor tissue, 1-124 FIAU activity decreases over
time, suggesting deiodination (Bengel et al. 2003). Studies employ-
ing F-18 FHBG demonstrate prolonged activity retention, rendering
this tracer approach superior for cardiac gene imaging. This may be
helpful in monitoring gene therapy in patients with end stage isch-
emic heart disease. In these patients, the coexpression of the thera-
peutic vascular endothelial growth factor (VEGF) gene (angiogen-
esis) as well as the imaging gene will allow unique assessment of
this new therapeutic approach.
1.3 Visualization of Target Proteins
It has been known for many years that tumor cells overexpress spe-
cific proteins, which can be targeted with peptides (Johnstrom et al.
2002). The most established peptide target in tumor cells is the so-
matostatin receptor, which represents a family of five subtypes (Reu-
bi et al. 1996). These receptors are primarily expressed in neuroen-
docrine tumors, small cell lung cancer, medullary thyroid carcinoma,
and lymphoma. Treatment of these cancers with somatostatin recep-
tor agonists has shown to have antiproliferative action. Using radi-
olabeled somatostatin analogs, specific visualization of receptors has
become possible (Froidevaux and Eberle 2002).
Refinement of the radio synthesis of somatostatin receptor antago-
nists by glycosylation of the radiopharmaceutical has yielded a high
Fig. 3. Visualization of somatostatin receptors overexpressed in neuroendo-

crine tumors. Comparison of SPECT imaging with In-Ill octreoscan (left)
and F-18-labeled receptor ligands. Please note the significantly improved
imaging provided by PET and fluor-labeled somatostatin receptor antagonists
tumor to non tumor contrast, which is far superior to the currently used
SPECT agent octreotide (Wester et al. 2003). Fluor-labeled gluco-
octreotate provides excellent imaging quality and allows identification
and localization of neuroendocrine tumors using whole body PET
imaging (Fig. 3).
A recently described target for imaging is adhesion proteins, such as
integrins, which are highly expressed in melanoma, breast tumor, os-
teosarcoma, and glioblastoma (Haubner et al. 2001). The scientific in-
terest of integrin imaging focuses on activated endothelial cells (an-
giogenesis) and cell migration. It is known that in the absence of oxy-
gen the activation of the hypoxia-inducing factor (HIF la) leads to
gene activation, which results in expression of VEGF, integrins, as
well as glucose transporters (GLUT1) and enzymes of the glycolytic
pathway. By targeting avfh, the integrin imaging should enable region-
Fig. 4. Autoradiographic measurement of radioactivity distribution in a RIP-

TAG tumor model of the mouse. In this model, tumor islands developed
within the pancreas, which are characterized by upregulation angiogenesis.
The areas of increased angiogenesis display increased binding of alphaV
beta3 ligands. Blocking studies demonstrate the specificity of this tracer re-
tention for avP3 integrins. (From Pichler et al. 2002)
al identification of areas with increased angiogenetic activity. Haubner

et al. of our group have succeeded in developing a radiotracer for the
identification of integrins (Haubner et al. 2001). Tumors overexpessing
a v /33 were labeled with fluor-18 RGD analogues with high tumor to
background ratio of tracer uptake, which can be blocked by cold antag-
onists indicating the specificity of the tracer signal. Results using cells
with various percentages of cells transfected with a v/33 genes suggest
that tracer techniques can be used to quantitate the expression of these
proteins. In a RIPl-Tag2 model, which is a transgenic mouse model of
pancreatic carcinogenesis and angiogenesis, the use of integrins al-
lowed visualization of the angiogenetic process, which was inhibited
by pretreatment of the animals with cold arginine-glycine-aspartic
acid (RGD) compounds (Fig. 4; Pichler et al. 2002).
These first studies indicate that imaging with specific peptides tar-
geting biological processes may be very helpful in the evaluation of
new therapies affecting, e.g., angiogenesis.
1.4 Assessment of Protein Function
A traditional application of tracer approaches is to measure protein

function in vivo. Radiolabeled substrates such as C-ll palmitate, C-
11 acetate, and F-18 deoxyglucose have been used for many years to
measure myocardial substrate metabolism. These tracer approaches
can be used to phenotype transgenic animal models in order to link
genetic profiles with measurement of protein function. In coopera-
tion with Dr. Charron - as an example for such an imaging
approach - GLUT-4 knockout mice, which lack the important insu-
lin-sensitive glucose transporter, are currently being investigated
(Simoes et al. 2003). These animals were studied prior to and after
insulin exposure. PET images revealed the lack of glucose uptake as
you would expect in the GLUT-4 knockout animals. However,
60 min after exposure to insulin there was no difference in FDG re-
tention in the knockout or the wild-type animals, indicating an adap-
tive process. Preliminary studies indicate that an increased expres-
sion of GLUT-l proteins may compensate for the loss of the GLUT-
4 transport capacity in the knockout animals, describing a new phe-
notype of glucose transport (Fig. 5).
Fig.S. PET studies of knockout mice (GLUT4) showing delayed response

to intravenous administration of insulin as compare to wild-type mice dis-
played in the upper row
1.5 Biological Imaging as a Clinical Approach
Besides the established use of radiolabeled glucose (FDG) in assess-

ing malignant transformed cells, several new imaging methodologies
have been introduced. Radiolabeled amino acids have been evaluated
by a variety of investigators (Weber et al. 1997, 2000; Langen et al.
2002). Especially in patients with brain tumors, radiolabeled amino
acids have proved to be superior to FDG (Weber et al. 1997). The
use of C-ll choline as a promising marker of tumor cell activity in
patients with prostate carcinoma has been advocated (Kotzerke et al.
2000). C-ll choline is not excreted by the kidneys and regional up-
take of the tracer in prostate cancer may allow sensitive identifica-
tion of malignant cells. However, especially in the lower abdomen
the correlation between morphology and function is important.
Therefore, choline imaging will particularly benefit from PET/CT in-
strumentation. Future studies will have to define the specificity of

this new tracer approach for malignancy of the prostate.
Aside from the basic metabolic activity of tumor tissue, the pro-
liferation rate is important in order to judge the aggressiveness of tu-
mor tissue. C-ll thymidine has been used for several years to mea-
sure the proliferation rate in tumor tissue, but the complex metabo-
lism of this compound makes the quantitative assessment of prolif-
eration difficult (Mankoff et al. 1999). Fluor-thymidine (FiT) has
been introduced by Shields et al. as an alternative to thymidine
(Shields et al. 1998). First clinical imaging studies show high FLT
uptake in proliferating tissue such as bone marrow. Studies correlat-
ing FLT uptake with proliferation rate measured in tumors with in
vitro techniques (Ki67) have indicated good agreement between both
parameters (Buck et al. 2002; Vesselle et al. 2002). Future therapy
studies will define the validity of such an approach, assessing the re-
sponse of tumor tissue to a given chemo- or radiotherapy.
Fig. 6. PETtCT fusion of atherosclerotic plaque imaging with F-l8 deoxy-

glucose. Please note the increased FDG uptake in the area of plaque as iden-
tified by contrast multislice CT imaging. (Used with permission of Circula-
tion, Rudd et al. 2002)
In terms of biological actlvIty, FDG tissue uptake has recently

been shown to be linked to acute and chronic inflammatory pro-
cesses. Large vessel arthritis is associated with intensive FDG up-
take, which disappears after successful therapy. Therefore, with the
improved spatial resolution of newer instrumentation, imaging of
vascular structures may become possible with biological tracers such
as FDG. This may be of utmost clinical importance in patients with
atherosclerotic disease, because inflammatory processes are linked to
the development of plaque ruptures. Figure 6 shows an example of a
PETtCT examination of a patient with carotid plaques. The fused
imaging clearly demonstrates high uptake of the tracer in the area of
the plaque formation.
1.6 Monitoring of Therapy Response
The most widely appreciated role of functional imaging focuses on

the monitoring of cancer therapy response. The need for imaging ap-
proaches to separate responders from nonresponders is highly appre-
ciated, especially among oncologists. Several studies have indicated
that the use of metabolic imaging allows assessment of the effect of
therapy based on FDG-images obtained prior to and after chemother-
apy (Schelling et al. 2000; Smith et al. 2000; Briicher et al. 2001;
Weber et al. 2001; Flamen et al. 2002; Weber et al. 2003). Correla-
tion of the FDG signals with histology of operative specimens after
therapy has indicated that PET can separate responders from nonre-
sponders based on quantitative measurements of FDG activity in the
tumor. Briicher et al. reported that tumor response can be assessed
based on a 50% reduction of the tracer signals with a 100% sensitiv-
ity and a 55% specificity in patients with esophageal cancer under-
going neoadjuvant chemotherapy (Briicher et al. 2001). The most ex-
tensive experience has been gathered in the area of lymphoma (Je-
rusalem et al. 1999; Spaepen et al. 2001; Spaepen et al. 2003). Pa-
tients who responded metabolically to therapy based on FDG signals
have a significantly longer survival rate compared to nonresponders.
More recent studies concentrate on the early definition of response.
Weber et al. were able to demonstrate, again in esophageal cancer,
that tumor response can be predicted as early as 2 weeks after initia-
tion of therapy. Weber et al. demonstrated that PET responders have

a significantly longer survival rate than PET nonresponders (Weber
et al. 2001). These important data imply that chemotherapy can be
stopped early in nonresponders and alternate therapy approaches se-
lected. On the other hand, PET responders may benefit from the
combined use of chemotherapy as well as surgical intervention.
1.7 Conclusion
There is no question that molecular imaging represents a rapidly ex-

panding field which will influence laboratory research as well as
clinical medicine. Due to the methodological differences of various
imaging approaches, multimodality imaging will be employed as a
very attractive evaluation of structure and function. PET/CT is the
first example, which combines the high spatial resolution provided
by helical CT with the high biological contrast provided by PET
imaging. The rapid clinical acceptance of this new imaging modali-
ty, especially in the United States, will produce the necessary clini-
cal validation to support further development of multimodality imag-
ing instrumentation.
In addition, imaging will be important for the preclinical and clin-
ical drug evaluation. Noninvasive strategies in the animal as well as
in patients will shorten the development process of new drugs and
reduce the need of large clinical studies if the correlation between
functional measurements and clinical outcome can be demonstrated
(surrogate endpoint). The first successful applications in patients
with lymphoma and gastrointestinal tumors will initiate imaging
strategies in the evaluation of new drugs. Imaging centers associated
with the pharmaceutical industry will emerge and integrate imaging
in the drug development process.
Finally, PET is currently the method of choice for biological
imaging due to its high sensitivity and its availability in the clinical
environment. The acceptance of PET imaging, especially in drug re-
search will depend on regulatory authorities allowing the prototyp-
ing of new radiopharrnaceuticals in order to enhance the drug evalu-
ation process. It is hoped that, with the close cooperation between
industry, universities, and regulatory agencies, PET will be an inte-
gral part in the drug evaluation process. This symposium is thought

to serve as an important platform to enhance the communication be-
tween these three parties and hopefully results in a coordinated strat-
egy to support the early use of biological imaging in drug develop-
ment.
Acknowledgements. The authors thank Gabriele Sonoda for her ex-

cellent secretarial support during completion of this manuscript.
References
Bell J, Taylor-Robinson S (2000) Assessing gene expression in vivo: mag-

netic resonance imaging and spectroscopy. Gene Ther 7:1259-1264
Bengel FM, Anton M, Avril N, Brill T, Nguyen N, Haubner R, Gleiter E,
Gansbacher B, Schwaiger M (2000) Uptake of radiolabeled 2' -fluoro-2'-
deoxy-5-iodo-l-j.l-D-arabinofuranosyluracil (PIAU) in cardiac cells after
adenoviral transfer of the herpesvirus thymidine kinase gene - the cellu-
lar basis for cardiac gene imaging. Circulation 102:948-950
Bengel FM, Anton M, Richter T, Simoes MV, Haubner R, Henke J, Erhardt
W, Reder S, Lehner T, Brandau W, Boekstegers P, Nekolla SG, Gansba-
cher B, Schwaiger M (2003) Noninvasive imaging of transgene expres-
sion using positron emission tomography in a pig model of myocardial
gene transfer. Circulation (in press)
Blasberg RG, Gelovani J (2002) Molecular-genetic imaging: a nuclear medi-
cine-based perspective. Mol Imaging 1:280-300
Bogdanov A, Weissleder R (1998) The development of in vivo imaging sys-
tems to study gene expression. Trends Biotechnol 16:5-10
Briicher BL, Weber W, Bauer M, Fink U, Avril N, Stein HJ, Werner M,
Zimmerman F, Siewert JR, Schwaiger M (2001) Neoadjuvant therapy of
esophageal squamous cell carcinoma: response evaluation by positron
emission tomography. Ann Surg 233:300-309
Buck AK, Schirrmeister H, Hetzel M, Von Der Heide M, Halter G, Glatting
G, Mattfeldt T, Liewald F, Reske SN, Neumaier B (2002) 3-deoxy-3-
[(l8)Flfluorothymidine-positron emission tomography for noninvasive as-
sessment of proliferation in pulmonary nodules. Cancer Res 62:3331-
3334
Chatziioannou A, Cherry S, Shao Y, Silverman R, Meadors K, Farquhar T,
Pedarsani M, Phelps M (1999) Performance evaluation of microPET: a
high-resolution lutetium oxyorthosilicate PET scanner for animal imag-
ing. J Nucl Med 40:1164-1175
Con tag P, Olomu I, Stevenson D, Contag C (1998) Bioluminescent indica-
tors in living mammals. Nat Med 4:245-247
Dai G, Levy 0, Carrasco N (1996) Cloning and characterization of the thy-

roid iodide transporter. Nature 379:21458-21463
de Vries E, Vaalburg W (2002) Positron emission tomography: measurement
of transgene expression. Methods 27:234-241
Doubrovin M, Ponomarev V, Beresten T, Balatoni J, Bornmann W, Finn R,
Humm J, Larson S, Sadelain M, Blasberg G, Gelovani T (2001) Imaging
transcriptional regulation of p53-dependent genes with positron emission
tomography in vivo. Proc Nat! Acad Sci USA 98:9300-9305
Flamen P, Van Cutsem E, Lerut A, Cambier J, Haustermans K, Bormans G,
De Lyn P, Van Raemdonck D, De Wever W, Ectors N, Maes A, Mortel-
mans L (2002) Positron emission tomography for assessment of the re-
sponse to induction chemotherapy in locally advanced esophageal cancer.
Ann Oncol 13:361-368
Froidevaux S, Eberle A (2002) Somatostatin analogs and radiopeptides in
cancer therapy. Biopolymers 66: 161-183
Gambhir S, Barrio J, Wu L, Iyer M, Namavari M, Satyamurthy N, Bauear
E, Parrish C, MacLaren D, Borghei A, Green L, Sharfstein S, Berk A,
Cherry S, Phelps M, Herschman H (1998) Imaging of adenoviral-directed
herpes simplex virus type 1 thymidine kinase reporter gene expression in
mice with radiolabeled ganciclovir. J Nucl Med 39:2003-2011
Haberkorn U, Altmann A (2001) Imaging methods in gene therapy of can-
cer. Curr Gene Ther 1: 163-182
Haubner R, Wester H, Weber W, Mang C, Ziegler S, Goodman S, Seneko-
witsch-Schmidtke R, Kessler H, Schwaiger M (2001) Noninvasive imag-
ing of alpha(v)beta3 integrin expression using F-18 labeled RGD-contain-
ing glycopeptide and positron emission tomography. Cancer Res
61:1781-1785
Jacobs A, Dittmar C, Winkeler A, Garlip G, Heiss W (2002) Molecular
imaging of gliomas. Mol Imaging 1:309-335
Jerusalem G, Beguin Y, Fassotte MF, Najjar F, Paulus P, Rigo P, Fillet G
(1999) Whole-body positron emission tomography using 18F-fluorodeoxy-
glucose for posttreatment evaluation in Hodgkin's disease and non-
Hodgkin's lymphoma has higher diagnostic and prognostic value than
classical computed tomography scan imaging. Blood 94:429-433
Johnstrom P, Harris N, Fryer T, Barret 0, Clark J, Pickard J, Davenport A
(2002) F-18-Endothelin-l, a positron emission tomography (PET) radioli-
gand for the endothelin receptor system: radiosynthesis and in vivo imag-
ing using microPET. Clin Sci (Lond) 103 Suppl 48:4S-8S
Knoess C, Siegel S, Smith A, Newpaort D, Richerzhagen N, Winkeler A, Ja-
cobs A, Goble R, Graf R, Wienhard K, Heiss W (2003) Performance
evaluation of the microPET R4 PET scanner for rodents. Eur J Nucl Med
Mol Imaging 30:737-747
Kotzerke J, Prang J, Neumaier B, Volkmer B, Guhlmann A, Kleinschmidt K,
Hautmann R, Reske SN (2000) Experience with carbon-II choline positron
emission tomography in prostate carcinoma. Eur J Nucl Med 27: 1415-1419
Langen KJ, Pauleit D, Coenen HH (2002) 3-[(123)I]Iodo-alpha-methyl-L-tyr-

osine: uptake mechanisms and clinical applications. Nucl Med Bioi
29:625-631
Lansford R, Bearman G, Fraser S (2001) Resolution of multiple green fluo-
rescent protein color variants and dyes using two-photon microscopy and
imaging spectroscopy. 1 Biomed Opt 6:311-318
Mankoff DA, Shields AF, Link 1M, Graham MM, Muzi M, Peterson LM,
Eary IF, Krohn KA (1999) Kinetic analysis of 2-[11C]thymidine PET
imaging studies: validation studies. 1 Nucl Med 40:614-624
Phelps M (2000) PET: the merging of biology and imaging into molecular
imaging. 1 Nucl Med 41:661-681
Pichler B, Braumuller H, Haubner R, Sakrauski A, Kneilling M, Seneko-
witsch-Schmidtke R, Schwaiger M, Roecken M, Weber W (2002) Moni-
toring of cellular immunotherapy in RIP1-Tag2 transgenic mice with
radiolabeled RGD-peptides. 1 Nucl Med 43: 122P (abstract)
Rudd lH, Warburton EA, Fryer TD, lones HA, Clark lC, Antoun N, lohn-
strom P, Davenport AP, Kirkpatrick Pl, Arch BN, Pickard lD, Weissberg
PL (2002) Imaging atherosclerotic plaque inflammation with [18F]-fluor-
odeoxyglucose positron emission tomography. Circulation 105:2708-2711
Reubi 1, Schaer 1, Laissue 1, Waser B (1996) Somatostatin receptors and
their subtypes in human tumors and in peri tumoral vessels. Metabolism
45(8 Suppl 1):39-41
Schelling M, Avril N, Nahrig 1, Kuhn W, Romer W, Sattler D, Werner M,
Dose 1, lanicke F, Graeff H, Schwaiger M (2000) Positron emission to-
mography using [(18)F]fluorodeoxyglucose for monitoring primary che-
motherapy in breast cancer. 1 Clin Oncol 18: 1689-1695
Shields AF, Grierson lR, Dohmen BM, Machulla Hl, Stayanoff lC, Law-
horn-Crews 1M, Obradovich JE, Muzik 0, Mangner TJ (1998) Imaging
proliferation in vivo with [F-18]FLT and positron emission tomography.
Nat Med 4:1334-1336
Simoes M, Egert S, Ziegler S, Miyagawa M, Reder S, Lehner T, Charron
M, Schwaiger M (2003) Delayed response of stimulated FDG uptake in
Glut4-null mice. 1 Nucl Med 44:190P (abstract)
Smith IC, Welch AE, Hutcheon AW, Miller JD, Payne S, Chilcott F, Waikar
S, Whitaker T, Ah-See AK, Eremin 0, Heys SD, Gilbert Fl, Sharp PF
(2000) Positron emission tomography using [(l8)F]-fluorodeoxy-D-glu-
cose to predict the pathologic response of breast cancer to primary che-
motherapy. 1 Clin Oncol 18:1676-1688
Spaepen K, Stroobants S, Dupont P, Van Steenweghen S, Thomas 1, Van-
denberghe P, Vanuytsel L, Borrnans G, Balzarini 1, De Wolf-Peeters C,
Mortelmans L, Verhoef G (2001) Prognostic value of positron emission
tomography (PET) with fluorine-18 fluorodeoxyglucose ([18F]FDG) after
first-line chemotherapy in non-Hodgkin's lymphoma: is [18F]FDG-PET a
valid alternative to conventional diagnostic methods? 1 Clin Oncol
19:414-419
Spaepen K, Stroobants S, Dupont P, Vandenberghe P, Maertens J, Bormans

G, Thomas J, Balzarini J, De Wolf-Peeters C, Mortelmans L, Verhoef G
(2003) Prognostic value of pretransplantation positron emission tomogra-
phy using fluorine 18-fluorodeoxyglucose in patients with aggressive
lymphoma treated with high-dose chemotherapy and stem cell transplan-
tation. Blood 102:53-59
Stegman L, Rehemtulla A, Beattie B, Kievit E, Lawrence T, Blasberg R,
Tjuvajev J, Ross B (1999) Noninvasive quantitation of cytosine deami-
nase transgene expression in human tumor xenografts with in vivo mag-
netic resonance spectroscopy. Proc Nat! Acad Sci USA 96:9821-9826
Tjuvajev J, Chen S, Joshi A, Joshi R, Guo Z, Balatoni J, Bailon D, Koutch-
er J, Finn R, Woo S, Blasberg R (1999) Imaging adenoviral-mediated
herpes virus thymidine kinase gene transfer and expression in vivo. Can-
cer Res 59:5186-5193
Tjuvajev J, Stockhammer G, Desai R, Uehara H, Watanabe K, Gansbacher
B, Blasberg R (1995) Imaging the expression of transfected genes in
vivo. Cancer Res 55:6126-6132
Vesselle H, Grierson J, Muzi M, Pugsley JM, Schmidt RA, Rabinowitz P,
Peterson LM, Vallieres E, Wood DE (2002) In vivo validation of
3'deoxy-3'-[(18)F]fluorothymidine ([(18)F]FLT) as a proliferation imag-
ing tracer in humans: correlation of [(18)F]FLT uptake by positron emis-
sion tomography with Ki-67 immunohistochemistry and flow cytometry
in human lung tumors. Clin Cancer Res 8:3315-3323
Weber W, Bartenstein P, GroB M, Kinzel D, Daschner H, Feldmann H, Rei-
del G, Ziegler S, Lumenta C, Molls M, Schwaiger M (1997) F-18-Fluoro-
deoxyglucose PET and I-I23-alpha-methyl-tyrosine SPECT in the evalua-
tion of brain tumors. J Nucl Med 38:802-808
Weber W, Wagner B, Anton M, Bodenstein C, Schwaiger M, Gansbacher B,
Haubner R (2003) Evaluation of the norepinephrine transporter as a re-
porter gene for monitoring of gene therapy. J Nucl Med 44:27P
Weber W, Wester H, Grosu A, Herz M, Dzewas B, Feldmann H, Molls M,
Stoecklin G, Schwaiger M (2000) O-(2-[18F]fluoroethyl)-L-tyrosine and
L-[methyl-ll C]methionine uptake in brain tumors: initial results of a
comparative PET study. Eur J Nucl Med 27:542-549
Weber WA, Ott K, Becker K, Ditt!er HJ, Heimberger H, Avril NE, Meiset-
schlager G, Busch R, Siewert JR, Schwaiger M, Fink U (2001) Prediction
of response to preoperative chemotherapy in adenocarcinomas of the eso-
phagogastric junction by metabolic imaging. J Clin Oncol 19:3058-3065
Weber WA, Petersen V, Schmidt B, Tyndale-Hines L, Link T, Peschel C,
Schwaiger M (2003) Positron emission tomography in non-small-cell
lung cancer: prediction of response to chemotherapy by quantitative as-
sessment of glucose use. J Clin Oncol 21:2651-2657
Weissleder R, Mahmood U (2001) Molecular imaging. Radiology 219:316-333
Wester H, Schottelius M, Scheidhauer K, Meisetschlaeger G, Herz M, Rau
F, Reubi J, Schwaiger M (2003) PET imaging of somatostatin receptors:
18 M. Schwaiger, W. Weber: Molecular Imaging: Dream or Reality?
design, synthesis and preclinical evaluation of a novel F-18-labeled, car-

bohydrated analogue of octreotide. Eur J Nucl Med Mol Imaging
30:117-122
Yaghoubi S, Barrio J, Dahlbom M, Iyer M, Namavari M, Satyamurthy N,
Goldman R, Herschman H, Phelps M, Gambhir S (2001) Human pharma-
cokinetic and dosimetry studies of F-18 FHBG: a reporter probe for
imaging herpes simplex virus type-l thymidine kinase reporter gene ex-
pression. J Nucl Med 42:1225-11234
Ziegler S, Pichler B, Boening G, Rafecas M, Pimps W, Lorenz E, Schmitz
N, Schwaiger M (2001) A prototype high-resolution animal positron to-
mograph with avalanche photodiode arrays and LSO crystals. Eur J Nucl
Med 28:136-143
2 Target Discovery and Validation
R. Solari
2.1 Introduction................................. 19
2.2 The Drug Discovery Process . . . . . . . . . . . . . . . . . . . . . . 21
2.3 Target Discovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.4 Target Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.5 Target Discovery Technologies . . . . . . . . . . . . . . . . . . . . . 26
2.5.1 Genomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.5.2 Bioinformatics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.5.3 Genetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.5.4 mRNA Expression Profiling . . . . . . . . . . . . . . . . . . . . . . . 33
2.5.5 Protein Expression Profiling. . . . . . . . . . . . . . . . . . . . . . . 36
2.5.6 Pathway Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.5.7 Metabolomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.5.8 Functional Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.6 Target Validation Technologies . . . . . . . . . . . . . . . . . . . . . 41
2.7 Conclusions................................. 44
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.1 Introduction
The pharmaceutical industry as we know it is still relatively young

and only started to emerge about 100 years ago, as advances in
chemistry and biology made the discovery and synthesis of drugs
possible. In the post-war years up to the late 1980s, the pharmaceuti-
cal industry grew at a spectacular rate and launched 44 new drugs in
one year at its peak in 1994. The success was based on a deep un-
derstanding of physiology and pharmacology coupled to brilliant
medicinal chemistry.
20 R. Solari
In the 1980s, the ability to rapidly clone and manipulate genes

heralded the start of the molecular biology era that culminated in
2001 with the publication of the first draft of the human genome se-
quence. This had an impact on the pharmaceutical industry that un-
derwent a dramatic revolution with the advent of the molecular phar-
macology era. The speed of the change is highlighted by the fact
that the first ,B-adrenergic drug was launched in 1965 and the cDNA
for the ,BTadrenergic receptor was only cloned as recently as 1986
(Dixon et al. 1986). These advances in molecular biology coincided
with breakthroughs in compound screening technology and chemical
synthesis. Engineering and robotics solutions allowed hundreds of
thousands of chemicals to be tested in biological assays every day,
so-called high-throughput screening (HTS). Chemistry also changed
to meet the growing demands for more compounds and more diver-
sity to fuel these screening engines and strategies for combinatorial
chemical synthesis evolved in this period. Thus, due to the introduc-
tion of high-throughput technologies in the 1980s, the drug discov-
ery process was changed from a methodical, empirical science to a
highly automated industry.
These dramatic changes to the drug discovery process were, and
still are, widely expected to improve productivity in the industry.
However, this has not yet materialised and there are some uncertain-
ties about the future direction of the industry. In 2001, only 25 new
drugs were launched, the lowest number for almost three decades.
The industry is spending ever-increasing amounts of money on re-
search and development, close to US $50 billion per year, and ac-
cording to the Tufts University Centre for the Study of Drug Devel-
opment (http://csdd.tufts.edu/), the cost of developing a new drug
has also risen from $231 million in 1987 to $802 million by 2000.
Investors expect major pharmaceutical companies to achieve 10%
annual growth, and in order to achieve this, each of the major
companies must launch, on average, four new chemical entities
(NCEs) per year with average sales of $350 million. However, from
1996 to 2001 the industry launched on average less than one NCE
per year per company and of all the drugs launched in 1996 only
25% had sales in excess of $350 million. Added to these pressures
are the looming patent expirations for many of the world's top sell-
ing drugs.
Target Discovery and Validation 21
Another significant trend in the late twentieth century is the emer-

gence of a biotechnology industry. Traditionally the pharmaceutical
industry made chemical drugs but molecular biology opened the way
for a new generation of biological drugs such as erythropoietin,
granulocyte colony-stimulating factor (G-CSF) and interferon-,B, a
number of which have now achieved blockbuster status in terms of
sales. The FDA (Food and Drugs Administration) approved the first
recombinant protein in 1982 and by 200 I biotechnology products
contributed 35% of new product launches. Many new vaccines and
monoclonal antibodies are currently in late stages of clinical trials
and represent a significant growth area for new biological medi-
Cllles.
2.2 The Drug Discovery Process
Going from a research idea to a marketed drug takes from 10 to

IS years and involves a vast range of technologies and specialities
along the way. Between 1990 and 1999, the pre-clinical phases took
3.8 years, the clinical trials 8.6 years and the US approval phase
1.8 years (DiMasi 2001). Time and failures are two key factors in
the process that successful companies work hard to reduce, and care-
ful historical analysis of each step has allowed the calculation of
failure rates in moving from one step to the next, often called the at-
trition rate. Attrition is so high that only about one in ten com-
pounds entering clinical trials will ever reach the marketplace, and
of these, only one in eight will recoup its investment costs. This
high failure rate continues back through every step in the discovery
process, and one finds approximately a 50% failure rate at each
stage. More detailed analysis has also shown that well-precedented
targets have lower failure rates than unprecedented targets. In this
context "precedented" means the target is well validated in the sci-
entific literature or, more importantly, drugs against this or similar
targets had been through phase II clinical trials or were already on
the market. A concern in the high-throughput post-genomic era is
that drug discovery pipelines will be filled with more unprecedented
or poorly validated targets, thereby increasing attrition rates and ulti-
mately costs. This fact has driven pharmaceutical companies to con-
22 R. Solari
centrate most of their efforts on target families that they believe will
have a higher chance of success. This leads us to the start of the
process - choosing a target.
2.3 Target Discovery
Most biological and medical research is still carried out in publicly

funded institutions, and curiosity driven research is still the main
source of innovation. Fortunately, much of this work is freely pub-
lished by the academic community and so provides many of the raw
materials for target selection. Reading the literature is clearly not en-
ough in a highly competitive environment, so how do pharmaceuti-
cal companies decide on which targets to start drug discovery pro-
grammes? Historically researchers studied physiology and pathology
and generated testable hypotheses on how to treat or modify certain
aspects of the disease. For example, understanding that adrenaline
has a central control over heart rate led to the development of adre-
nergic receptor antagonists (beta blockers) for the control of heart
disease. This experimental strategy and the joint efforts of biologists
and chemists generated most drugs on the market today. In the
"post-genome" era, the way that biology research is conducted has
changed. The first and most obvious change is that we now have a
list of all of the possible drug targets in the human body. The new
challenge is to pick the right ones from this list, a process that can
be thought of as "reverse biology".
One of the most useful techniques for target discovery is bioinfor-
matics. This term describes a broad range of computational tech-
niques that seek to add biological significance to raw DNA sequence
data. By recognising characteristic motifs or domains in protein se-
quence, it is sometimes possible to assign a putative function that
can be subsequently validated by experimentation.
However, genomics and bioinformatics alone are still not enough
to choose a drug target. There are many hundreds of interesting can-
didates in the genome and it would not be feasible to start drug dis-
covery programmes on each and everyone. Consequently a process
of target selection has to be performed and this can be done in a
number of ways. One of the simplest ways is to examine levels of
expression of a particular protein or its mRNA in healthy versus dis-

eased tissues, techniques known as proteomics and differential gene
expression (DGE), respectively. This gives a first clue that a particu-
lar protein may be implicated in the disease process, although it
does not distinguish cause and effect. An alternative approach is to
use genetics. Many diseases have a heritable component, and study-
ing the inheritance of these traits can identify genes that cause or
predispose the individual to that disease. There are many genetic
strategies that can be adopted that range from positional cloning of
candidate genes as they are inherited through family pedigrees to as-
sociation genetics which is the study of the linkage of certain genet-
ic loci with disease incidence across large populations. Whereas the
link between the gene and the disease can be relatively clear in cer-
tain monogenetic diseases such as cystic fibrosis, in more complex
multifactorial diseases, how a particular allele contributes to the
pathology may be difficult to understand or translate into an obvious
drug target. Sometimes a great deal of biological research has to be
done following the identification of a candidate gene in order to
map a particular pathway and to place the candidate into an appro-
priate context. Invertebrate model organisms such as yeast, nema-
todes and fruit flies are particularly tractable by genetic means and
are useful tools for mapping pathways. Pathway mapping can also
be performed by identification of protein - protein interaction net-
works using techniques such as yeast-two-hybrid screening (Uetz et
al. 2000; Ito et al. 2001) or by purification of protein complexes and
identification of all of the components by proteomics techniques
(Gavin et al. 2002; Ho et al. 2002). Mouse genetics (Zhang et al.
1994) and more recently zebrafish have also been valuable models
for identification of disease candidate genes. Knocking genes out,
either specifically or by random mutagenesis, followed by careful
phenotypic analysis can often lead to the identification of novel drug
target candidates.
These various techniques generate a candidate drug target that is
hopefully free from predating patent filings and that needs to be va-
lidated before progressing along the drug discovery process.
24 R. Solari
2.4 Target Validation
Drugs can fail for a variety of reasons, one of which is that the ini-
tial target chosen for the drug discovery is inappropriate. Many peo-
ple in the industry will argue that the only truly validated targets are
those for which there is a drug already in man. For the purposes of
progressing a candidate target into drug screening, validation has an-
other perhaps less stringent meaning - generating as much confi-
dence as possible that a drug against the target will modify the dis-
ease or treat the symptoms. Initially a candidate may have relatively
little scientific weight. For example it may be a novel G protein-
coupled receptor (GPCR) that is expressed in a particular region of
the brain or a novel cell surface protein over-expressed on the sur-
face of a particular cancer cell, and this data may have come from
mRNA expression profiling. First, one would want to confirm the
observation across multiple independent samples, and then one
would want to confirm that the protein levels are also differentially
expressed. At this point one has a correlation between protein ex-
pression and some observed physiology or pathophysiology. The
challenge is then to show that the expression of this particular target
is either causative of the disease or some of the symptoms of the
disease. Placing a candidate in the context of a disease requires the
use of model systems. These can be cell or tissue based or whole
animals, but in whichever case they must attempt to be predictive of
the situation in man.
There are many available cell lines where some relevant phenoty-
pic readout can be conveniently assayed. Candidate protein targets
can be over-expressed in these cells lines by transfection of the cor-
responding cDNA and the effect on the phenotype measured. Simi-
larly, a candidate target protein can be "knocked out" by expression
of a dominant negative mutant version of the protein or by inhibit-
ing protein expression with antisense oligonucleotides or small inter-
fering (si)RNA. These cell-based techniques can give increased con-
fidence that stimulating or inhibiting a particular target or target
pathway may have therapeutic utility and represent a major tool for
target validation. These techniques are also scaleable and a number
of such "high-throughput biology" strategies are being developed
with the aim of validating targets in a genome-wide fashion, one ob-
vious example being genome-wide use of siRNA (Gonczy et al.

2000).
Animal models, such as knock-out or transgenic mice, are ex-
tremely valuable in target validation. The hope is that by modifica-
tion of the target expression or by expression of a mutant allele,
some aspects of the human pathology will be recapitulated in the
mouse.
The aim of the target validation step is to generate enough confi-
dence in the target that the expensive and time-consuming process
of drug discovery should begin. The better the validation of the bio-
logical hypothesis the less the chance of failure further downstream.
Thus the output from the target validation step should address the
following questions:
1. Is there sufficient evidence that the target is implicated in disease

pathology?
2. Is there evidence that activation or inhibition of the target's activ-
ity would modify the disease or provide symptomatic relief?
3. Is the target chemically tractable?
4. Is the target tractable with a biological agent?
5. Is it possible to configure an appropriate screen to identify agents
that activate or inhibit the target?
If all of these scientific criteria are met, there usually follows a

more clinical and commercial analysis at this stage. If one had a
drug that acted on this target, how would one configure a clinical
trial, how would it be used clinically, what are the current gold stan-
dard therapies and how would this compare? It may seem preco-
cious to consider such distant issues so early in the drug discovery
process; however, experience tells us that if these issues are not
clear at the outset, the project may fail later in the process when
much more time and money has been spent. This process is usually
known as defining the product profile.
26 R. Solari
2.5 Target Discovery Technologies
2.5.1 Genomics
On 14 April 2003, at the 50th anniversary of the discovery of the

structure of DNA, the Human Genome Sequencing Consortium an-
nounced that it had completed its work (http://www.sanger.ac.ukJ).
In June 2000 both the public consortium and Celera Genomics an-
nounced draft sequences and both published their reports simulta-
neously in February 2001 (International Human Genome Sequencing
Consortium 2001; Venter et al. 2001). The new reference sequence
is more complete and accurate than the draft sequences with an er-
ror rate of about one in 100,000 bases and all but a few remaining
repetitive stretches and the centromeres are now finished. Neverthe-
less, this astonishing achievement represents the start rather than the
finish. We still have not determined the total number of genes in the
human genome. Initial estimates put the number at 35,000 to 45,000
and current thinking is that it may be closer to 30,000, and as many
as half of these genes still have largely unknown functions. There is
no doubt that the genome sequence has changed the way biomedical
research is performed, and there is also no doubt about its massive po-
tential value; however, there is still controversy as to whether it has
made the process of drug discovery more efficient, and there is a grow-
ing desire to tum this genome sequence into real tangible benefits.
The human is not the only species to have its genome sequenced.
Many model organisms including the mouse, zebrafish, fruit fly,
nematode and yeast have all been sequenced. These organisms are
genetically tractable in the laboratory and have been the favourite
model systems for geneticists trying to unravel gene function. Hav-
ing access to their genomic sequence has accelerated and facilitated
the use of these genetic models. The genomes of many pathogens
such as viruses, bacteria and protozoa have also been sequenced.
This is perhaps where there is the greatest hope of translating geno-
mic information directly into therapeutic benefit. Within weeks of
the outbreak of the severe acute respiratory syndrome (SARS) epi-
demic in Southeast Asia, scientists reported the genome sequencing
of the suspected coronavirus pathogen. This is a reflection that we
have truly entered the era of genomic medicine.
2.5.2 Bioinformatics
Translating this mass of DNA sequence data into useful information

is a daunting challenge. This DNA sequence data is of two forms,
genomic and cDNA sequences. Since cDNA is made from mRNA,
the sequence information defines a transcribed gene. Genomic DNA,
by contrast, is made up of introns and exons and vast stretches of
untranscribed DNA, and hunting for genes in this mass of sequence
is not trivial. There is a wide range of computational techniques
available for searching through genomic sequences to search for
genes. A major public effort is a joint project between the European
Molecular Biology Laboratory (EMBL-EBI) and the Sanger Centre
called Ensembl (www.ensembl.org). Ensembl automatically tracks
all genomic sequence data from the human genome and attempts to
assemble the DNA sequences into single contiguous stretches and to
then analyse the sequences for features of biological interest, usually
known as "annotation". The annotation of the genome firstly com-
prises the genes themselves. These can be either predicted by En-
sembi software or be previously known from published cDNA se-
quencing efforts. The annotation also identifies common single nu-
cleotide polymorphisms (SNPs) that are single base pair changes in
the genome that are found between different individuals and occur
with a frequency of once every 100-300 base pairs. These SNPs
provide a high-resolution map of the genome that is very useful for
tracking the inheritance of various stretches of DNA from one gen-
eration to another.
Bioinformatics is not only concerned with predicting genes but
also with trying to predict the function of gene products, that is pro-
teins. Thanks to the many years of accumulated knowledge of mo-
lecular cell biologists and biochemists, we can define domains of
proteins that are responsible for particular functions. For example,
we now know the archetypal amino acid sequence that defines a pro-
tein kinase domain. Many such functional domains have been de-
fined such as the signal peptide for secreted proteins and the cataly-
tic domains of enzymes. These signature amino acid sequences can
be used to search the genome databases to identify other genes re-
lated to the functional domain of interest. For example, one can
search the genome to identify all genes with a sequence related to
28 R. Solari
the protein kinase domain. In this way, one can identify approxi-
mately 500 kinases in the human genome, some of which were pre-
viously known and some of which are novel. The members of any
gene family will show various degrees of relatedness to the archety-
pal signature sequence and a great deal of mathematical analysis is
performed to assess if a particular gene is really a member of a fam-
ily of not.
Bioinformatics also deals with predicting protein structure. The
amino acid sequence is known as the primary structure, and the lo-
cal folding of the protein chain into helices or sheets is known as
the secondary structure. There are now many computational tech-
niques for secondary structure predictions based on homology to
known structures of related gene family members or based on de
novo prediction methods. Protein fold predictions are another way to
try and define protein function and so to "annotate" the genome.
How does genomics and bioinformatics impact the drug discovery
process? Perhaps the most dramatic example is the discovery of cy-
clo-oxygenase (COX)-2, which led to the subsequent development of
COX-2-selective drugs that now represent a multi-billion dollar mar-
ket. In 1991, workers studying early response genes in fibroblasts
transformed with Rous sarcoma virus identified an mRNA transcript
coding for a protein with sequency similarity to seminal vesicle
COX (Xie et al. 1991). An independent study also identified a novel
transcript whose sequence was related but not identical to COX
when studying phorbol ester-induced genes in 3T3 cells (Kujubu et
al. 1991). These novel COX-related transcripts turned out to be
COX-2. Recently, it has been shown that isoforms of COX-l exist,
called COX-3, and these findings are set to have a profound influ-
ence on our understanding of COX pharmacology (Chandrasekharan
et al. 2002).
Searching the genome for novel family members of well-prece-
dented targets, such as COX, clearly can be very profitable. Several
groups have taken a systematic view of this problem and tried to
analyse the protein targets of all successful chemical medicines to
date. An initial survey suggested that these drugs had targeted ap-
proximately 500 proteins (Drews 1996; Drews and Ryser 1997).
This led to the now widely held belief that only a particular set of
proteins are able to be acted on by drugs, and these proteins are of-
ten termed "druggable" or "tractable". This may reflect that most

small molecule drugs act by competing with an endogenous small
molecule ligand for a privileged binding pocket. The fact that the
"druggable" genome is relatively limited is that there are only a lim-
ited number of such small molecule binding pockets. Extrapolating
this across the whole genome suggested that there might be as many
as 5,000 to 10,000 of these "druggable" or "tractable" targets. More
recent estimates suggest that if one considers only Lipinski-compli-
ant agents, only 120 proteins are the targets of all of today's mar-
keted drugs, and the total druggable genome is represented by 3,051
targets (Hopkins and Groom 2002). However, of these 3,051, many
would not have any influence on disease and perhaps only 600-
1,500 may actually represent valid drug targets. The development of
bioinformatics tools allowed researchers to place these known tar-
gets into families and it is now accepted that certain protein families
are the most favourable targets for small molecule drug discovery.
These include some classes of GPCRs, ion channels, nuclear recep-
tors and enzymes. It is fair to say that most major pharmaceutical
companies now focus most, if not all, of their small molecule drug
discovery efforts on these favoured druggable targets. However,
bioinformatics can also be used to mine the genome for proteins
with a leader sequence. Such proteins may be secreted growth fac-
tors or hormones and thus may have therapeutic utility themselves
or may be cell surface antigens and so represent potential targets for
antibody therapeutics. Many biotechnology companies are exploiting
targets such as these for the development of biological drugs. Thus,
genomics coupled with bioinformatics without any deeper biological
experimentation can generate a list of target "candidates". However,
"reverse biology" strategies based on genomics and bioinformatics
can only take research a short distance towards the goal of target
discovery.
2.5.3 Genetics
The study of heritable traits has long provided a rich source of bio-
logical knowledge, particularly in model organisms that are amen-
able to genetic manipulation in the laboratory. The favoured models
30 R. Solari
have been the yeast (Saccharomyces cerevisiae, http://genome-

www.stanford.edu/Saccharomycesl), the nematode worm (Caeno-
rhabditis elegans, http://www.wormbase.org/), the fruit fly (Droso-
phila melanogaster, http://fIybase.bio.indiana.edul), the mouse
(http://www.informatics.jax.org/), and more recently the zebrafish
(Danio rerio, http://zfin.org/zLinfo/dbase/db.html). Either by selec-
tive breeding or by deliberate mutagenesis varieties of these species
could be generated that displayed a particular physical characteristic
or phenotype. Using a variety of molecular biology techniques it is
possible to identify the genetic alteration that gave rise to the partic-
ular characteristic - thus linking the gene to the phenotype. How-
ever, the phenotype of an organism is a result of many gene prod-
ucts interacting and model organism genetics is a powerful tech-
nique for discovering such genetic interactions. As most biochemical
events require proteins to act in a defined sequence along a pathway
to produce an end product, mutations in any of the genes encoding
those proteins can give rise to the same phenotype. One can exam-
ine the interaction of genes by crossing these mutant organisms and
determine whether the genes act on the same pathway by seeing if
by their combination they enhance or suppress the observed pheno-
type. A gene or locus that suppresses or masks the action of a gene
at another locus is said to be epistatic. The elucidation of biochem-
ical pathways by epistatic analysis in models such as C. elegans and
D. melanogaster has had significant impact on human biology be-
cause many pathways are highly conserved throughout evolution.
The regulation of the cell cycle, ras signal transduction, and the nu-
clear factor (NF)-KB pathways are just a few of the many complex
cellular processes that have been elucidated in simple model organ-
Isms.
Genomics has also had a major impact on genetics. Genetics tra-
ditionally relied on random mutagenesis followed by tracking down
the mutant gene of interest. Since the complete genomic sequence of
these model organisms is now known, it is possible to systematically
knock out every gene in the genome one at a time and directly in-
vestigate the association of each gene to a phenotype. There are a
number of large-scale public and private efforts to systematically
knock out every gene and to assign a phenotypic consequence (see
http://www.biocat.de/open_biosystems_YeastKO.htm). New technolo-
gies are also making an impact and genome-wide knock-out strate-

gies using siRNA now appear to be possible (Gonczy et al. 2000).
Although the yeast, nematode and fruit fly have made massive
contributions to our understanding of biological systems, the most
studied vertebrate is the mouse. Although the mouse is less experi-
mentally convenient than these invertebrate models, it clearly is pre-
ferable for studies on more complex systems such as brain function,
immunology, cardiovascular and respiratory biology. Mouse knock-
out or transgenic technology is far from perfect, and many mutants
either have no observable phenotype or are lethal. However, a retro-
spective analysis of published gene knock-out studies in the mouse
of the 100 top drug targets revealed that in most cases the pheno-
type correlated quite well with the drug efficacy (Zambrowicz and
Sands 2003). Mouse knock-out strategies can either be highly direc-
ted or random. In the case of directed knock-outs, the specific gene
of interest is knocked out by inserting a selectable marker gene in
the place of the targeted gene. In the biopharmaceutical industry
there is a major effort to systematically knock out and analyse the
phenotype of all the "druggable" genes in the genome. In the ran-
dom approach, mice are subjected to mutagenic chemicals such as
N-ethyl-N-nitrosourea (ENU) and large breeding colonies are estab-
lished. By careful monitoring of the mice, mutant phenotypes can be
detected and by conventional genetics and molecular biology the ge-
netic mutation that gave rise to the altered phenotype can be tracked
down. Mouse genetics has a long history and there are several orga-
nisations that have been breeding mouse varieties for decades. Many
of the varieties that have been bred have provided major insights
into human pathology such as the obese mouse that led to the dis-
covery of leptin (Zhang et al. 1994).
Ultimately, one is trying to make medicines for human diseases,
so the most appropriate model is man. Genetic manipulation of man
is neither ethically acceptable nor practically possible; however, we
are increasingly studying the genetic basis of human disease. There
are a number of diseases where mutations in a single gene result in
the inheritance of a particular disease condition. There are some
well-known examples of these relatively rare diseases such as cystic
fibrosis. By following the inheritance of the disease condition
through family pedigrees it is possible to identify genetic markers
32 R. Solari
that co-segregate with the disease gene, showing that they are physi-
cally linked, that is, they are on the same chromosome. The statisti-
cal frequency that the genetic marker co-segregates with the disease
gene is related to how close the disease gene is to the genetic mar-
ker. By continually refining the analysis it is possible to define a
smaller and smaller region of the genome in which the disease gene
is found until it is finally identified. This approach is called posi-
tional cloning and was used to identify the genes mutated in cystic
fibrosis and a familial form of breast cancer (BRCAI). This strategy
works well for diseases caused by mutations in a single gene; how-
ever, most of the common human diseases such as asthma, rheuma-
toid arthritis, cardiovascular disease and cancer are multifactorial
and result from a combination of many genes and environmental fac-
tors. No single gene or influence is usually sufficiently penetrant for
it to be accurately followed through family pedigrees, and an alter-
native genetic approach is usually needed, called association genet-
ics. In these studies, large numbers of affected and unaffected indivi-
duals are collected from a population, and the association of genetic
markers with the incidence of the disease are identified. These ge-
netic markers, these days usually SNPs, help to identify regions of
the genome that are statistically significantly co-inherited with an in-
creased susceptibility to the disease. By examination of the whole
genome sequence it may be possible to identify interesting candidate
genes that lie within this region, and from there it may be possible
to construct a testable hypothesis to examine the role of that gene in
the development of the disease. When compared to linkage analysis,
association analysis offers superior statistical power for detection of
genes that manifest a modest phenotypic effect.
Whether a disease candidate gene has been identified by posi-
tional cloning, as in a simple genetic diseases such as cystic fibrosis,
or whether the candidate has been found as a disease susceptibility
gene by association genetics, there may not be a clear therapeutic
strategy for the treatment of the disease. The candidate gene may
clearly identify the molecular cause of the disease, but the gene
product may not be amenable to a therapeutic approach simply be-
cause it does not belong to a druggable class of target. The candi-
date gene may encode a protein of unknown function, as was the
case with BRCA 1, and a great deal of biology has to be done to
generate a testable hypothesis. In polygenic diseases, the candidate

genes also may not be druggable and moreover their contribution to
the final disease pathology may be far from clear. Perhaps the best
example of the role of genetics for target discovery in a complex
disease is in Alzheimer's disease. In about 5% of cases the disease
follows a highly penetrant familial inheritance, whereas in the ma-
jority of cases the disease appears to have a sporadic incidence. Ge-
netic association studies revealed that development of early onset
Alzheimer's disease was strongly associated with the inheritance of
the ApoE4 allele. However, the biological rationale for how ApoE4
predisposes to the development of the disease remains unclear and it
has so far proved to be an unpromising avenue for drug discovery.
The rare familial form of the disease is associated with mutations in
two proteins, the amyloid precursor protein (APP) and presenilin
(PS). The biology of these two proteins started to provide an attrac-
tive mechanistic hypothesis to explain the formation of amyloid pla-
ques in the brain. PS turned out to be one of the two proteases re-
quired to cleave APP to generate the amyloidogenic AfJ42 peptide.
The other protease, fJ-secretase or BASE was subsequently discov-
ered by a variety of biochemical and molecular biology techniques,
and because these proteases are druggable targets many companies
are now actively engaged in the development of inhibitors for the
treatment of Alzheimer's disease (De Strooper and Konig 1999).
Time will tell if the APP/PS pathway is more or less important than
ApoE4 in the development of Alzheimer's drugs; however, in terms
of drug discovery there is a pragmatic view that you work with sys-
tems that you can most readily test.
2.5.4 mRNA Expression Profiling
Genomics coupled with bioinformatics is capable of generating a list

of the potential drug targets. As discussed previously, most pharma-
ceutical companies are currently focused on GPCRs, ion channels,
nuclear receptors, kinases and a variety of enzymes families. How-
ever, these "potential" targets have to be placed in the context of
biology and pathology. The simplest way of getting some initial in-
formation about what a protein is doing is to identify in which tis-
34 R. Solari
sue it is expressed. There are a number of public and private banks

or repositories of human tissue and there has been a considerable ef-
fort to determine the tissue distribution of the major druggable gene
families. This can be done by a variety of techniques including PCR
(polymerase chain reaction) or hybridisation methods to detect
mRNA in biopsies or on tissue sections or protein detection using
immunohistochemistry (IHC). Expression of an mRNA in a particu-
lar tissue does not mean that the encoded protein represents a drug
target, a great deal of biological and pharmacological validation still
needs to be done; however, it is a starting point.
The next level of mRNA expression profiling is to compare
healthy with diseased tissue to try and define the observed pathology
at a molecular level. If a gene is differentially expressed in a dis-
eased compared to a healthy tissue, then it may represent a drug tar-
get. However, the contrary is not true. If a gene is not differentially
expressed in a disease, this does not mean that the gene product is
not a drug target. Most successful drugs act on targets that are not
directly involved with the disease pathology but rather treat symp-
toms.
The technology for rapid and comprehensive mRNA expression
profiling using microarrays has become particularly well advanced
in recent years. Microarray analysis has become a widespread tool
for target discovery and validation as well as a platform for diagnos-
tics. The technology is based upon extraction of mRNA from cells
or tissues, labelling it and hybridisation to high density arrays of oli-
gonucleotides or cDNA on solid surfaces such as glass slides or fil-
ters. The array elements bind to their corresponding mRNA and give
a quantitative measure of abundance for each mRNA species in the
mixture. The advantages of these techniques over conventional mo-
lecular biology techniques such as Northern blotting include the fact
that tens of thousands of mRNAs can be measured simultaneously,
rapidly and at relatively low cost. The huge amount of data gener-
ated require sophisticated data analysis and modelling software.
There are two fundamental platforms in common use. The first uses
oligonucleotide arrays (Affymetrix) and the second uses arrays of
PCR-amplified cDNA (100-3,000 base pairs). For oligonucleotide
arrays, mRNA is purified from the two samples to be compared and
labelled with a fluorescent dye prior to hybridisation to two array
chips. The quantitative signals from the two chips can be electroni-
cally compared to identify changes in gene expression between the
two samples. For cDNA arrays, the two mRNA samples are labelled
separately with two different coloured fluorescent dyes, usually cya-
nine 3 and cyanine 5 (Cy3 and Cy5), then mixed and hybridised to
a single microarray slide. The slide is analysed for fluorescent inten-
sity in the two colour channels and an overlay of the data is used to
identify changes in mRNA abundance between the two samples.
The main alternative to microarrays for gene expression profiling is
a technique called SAGE (serial analysis of gene expression) http://
www.sagenet.org/.This method purifies mixtures of mRNA from bio-
logical samples then cuts it into pieces using specific enzymes that
generate short tags of 10-14 base pairs. The collection of tags are
joined together into one long piece of DNA that is then sequenced
and the frequency of each tag is determined. The frequency of each
tag is related to the abundance of the parent mRNA in the original
sample and the tags are sufficiently long so that the identity of the par-
ent mRNA can be determined by database analysis. Thus SAGE is an
"open" platform that will detect all changes in mRNA expression,
whereas microarrays will only detect changes in those mRNA species
for which a specific probe has been deposited on the array.
There is no denying the success of mRNA expression profiling as
a technology platform, judging by the enthusiasm with which it has
been adopted by the scientific community; however, it is reasonable
and perhaps timely to question whether this technology has made a
significant impact on biomedical research. This technique generates
unprecedented quantities of data, many millions of data points, and
biologists on the whole are poorly equipped to analyse and make
sense of such data. In addition, when generating large data sets such
as these, the original study design is of crucial importance and
should be based on sound biological and mathematical footings. All
too often this is not the case. Finally, there are a number of different
technology platforms in use today that are difficult to compare one
to the other. Reproducibility and comparability of the data between
laboratories makes interpretation very difficult indeed. Collecting
data has never been easier for biologists; however, less thought has
gone into turning this data into valuable knowledge. Nevertheless,
microarray studies have made some significant contributions, partic-
36 R. Solari
ularly in the field of oncology, where a number of studies have

shown that it is possible to subdivide tumours by mRNA profiling
that distinguishes them based on progression and responsiveness to
therapeutic regimes (Bhattacharjee et al. 2001).
2.5.5 Protein Expression Profiling
Measuring changes in cellular mRNA expression can provide valu-

able information about how a cell responds to a particular stimulus
or set of stimuli. However, changes in mRNA do not accurately in-
form about how protein levels change in the cell, which is ulti-
mately what controls the cellular response. Consequently, measuring
protein changes is one step closer than mRNA profiling for relating
the genome to a biological response, and there is much effort going
into protein expression profiling or "proteomics". The most common
technology platform is to extract proteins from cells, tissues or bio-
logical fluids and to analyse these complex mixtures by two-dimen-
sional polyacrylamide gel electrophoresis (2D-PAGE). The pattern of
expressed proteins on the 2D-PAGE can be analysed by image anal-
ysis software to identify changes between one sample and another.
Protein spots of interest can be cut out of the 2D-PAGE, enzymati-
cally digested and the protein identified by a variety of mass spec-
trometry (MS) techniques. The promise of proteomics is very excit-
ing; however, there are both technical and intellectual issues that
may ultimately limit its impact. The first limitation is the 2D-PAGE
technology itself. Running gels is difficult to scale, tedious and suf-
fers from poor reproducibility. Certain classes of proteins, such as
hydrophobic membrane spanning proteins or basic nuclear proteins,
are underrepresented in the gels and consequently are largely missed
in the analysis. The technique is biased towards abundant proteins
and special methods are required to detect rare proteins. Finally, the
complexity of protein post-translational modifications is daunting.
So far, there are more than 100 modifications that have been de-
scribed, such as phosphorylation, glycosylation, sulphation and so
on. These modifications may produce millions of chemically distinct
protein entities within the cell and every cell type may be different.
Thus the proteome is a variable, context-specific collection of pro-
teins unlike the genome, which is a relatively fixed entity. This bio-
logical complexity and the technical difficulties make protein expres-
sion profiling a daunting, but not impossible challenge.
One solution is a process known as SELDI (surface enhanced la-
ser desorption ionisation mass spectrometry) which has done away
with the gel technology altogether. Biological samples are applied to
chromatography chip surfaces that capture some of the proteins from
the mixture. These chips can simply be chemically modified to
make them hydrophobic or charged, or they can be coated with a
biological agent such as an antibody, an enzyme or a receptor. The
captured proteins are then "read" by excitation with a laser and anal-
ysis of protein mass by MS to produce a mass signature. These sig-
natures can rapidly be compared between multiple samples to deter-
mine changes in protein expression.
As an alternative to MS-based technologies for proteomics, there
have been attempts to develop protein microarrays. The basic idea is
to array large numbers of proteins onto a chip and to incubate the
chip with a biological sample. If there are components in the biolog-
ical sample that bind to proteins on the array, they can be quantified
using an appropriate detection assay. One possible form of such a
protein chip might be an array of antibodies, and ultimately one
may imagine that such protein chips could be constructed with ge-
nome-wide coverage. Such an antibody chip would provide a protein
expression profiling platform. One example of a genome-wide pro-
tein chip already exists for yeast (Zhu et al. 2001). In this study,
5800 individual yeast proteins (93.5% of the whole genome) were
expressed, purified and printed in an array on glass slides. The array
was then probed with biotinylated calmodulin to identify calmodu-
lin-binding proteins and with phosphoinositide (PI) containing lipo-
somes to identify PI binding proteins. This study demonstrated that
genome-wide protein chips are a real possibility.
2.5.6 Pathway Mapping
Measuring changes in individual proteins or mRNAs provides a low

resolution view of the cellular or tissue response to stimuli. How-
ever, biological functions require proteins to act in multi component
38 R. Solari
machines and to work in pathways to produce an end product. In or-

der to more fully appreciate whether or not a protein is a drug target
one needs to place it in the context of its biological function. As de-
scribed above, model organism genetics is particularly powerful at
mapping biochemical pathways; however, proteomics and genomics
techniques are starting to make major contributions in this field, giv-
en the fact that they can be made scaleable and high throughput.
These strategies are based upon the identification of large-scale pro-
tein-protein interaction maps to define all the protein complexes in
the cell. The rationale is that the function of an unknown protein
can to some extent be defined by identification of the other proteins
that it binds to. It is estimated that all proteins interact with between
2-10 other proteins, and if all protein-protein interactions can be de-
fined, then it would provide a basic framework or road map for how
cells are organised.
One strategy is to use the yeast-2-hybrid (Y2H) technique and
two large-scale studies have been published that have sought to de-
fine comprehensive protein interaction maps for the budding yeast
(Uetz et al. 2000; Ito et al. 2001). The Y2H assay is based on the
observation that Ga14 transcription factor in yeast can be divided
into two functional domains, a DNA-binding domain (BD) and an
activation domain (AD). If expressed as individual domains, they are
unable to drive transcription from a Ga14 promoter; however, if the
two separated domains can be forced to bind together then transcrip-
tion is activated. By fusing a "bait" protein (protein X) to the BD
and a "prey" protein (protein Y) to the AD, one can test if X will
bind to Y. If X and Y bind to one another they will bring the BD
and the AD together and the combination of the two fusion proteins
will activate transcription of a reporter gene from a Ga14 promoter.
The reporter gene can be one that allows selection of the yeast cells
in which X and Y bind, so allowing the assay to be run as a high-
throughput screen. In this way a bait of protein X can be screened
against a library of potential prey proteins fused to the AD to identi-
fy potential interacting partners. In the published large-scale studies,
most of the open reading frames (orf) in the yeast genome were in-
dividually fused to the Ga14 DNA binding domain, and each of
these baits was screened against a library of yeast orfs fused to the
Ga14 activation domain. This strategy should identify all of the pair-
wise protein-protein interactions in the yeast cell. Interestingly the

two studies identified similar numbers of interactions; however, there
was a surprisingly low level of overlap between the proteins identi-
fied and between the protein interactions that had already been de-
scribed in the literature. Clearly, large-scale Y2H studies can be per-
formed in high throughput but there are serious issues to be exam-
ined relating to the reliability of the data sets generated.
An alternative approach to producing large-scale protein-protein
interaction maps is to use protein biochemistry to purify protein
complexes and MS techniques to identify components of the com-
plex. The strategy here is to immunoprecipitate a target protein of
interest under conditions that retain binding of associated proteins in
a protein complex. A variety of techniques have been used, but the
most common involve introducing a stretch of DNA encoding a
short peptide tag onto a cDNA target of interest that is then over-ex-
pressed in an appropriate cell. The cell extract is then immunopreci-
pitated with an anti-tag antibody, and the purified target and asso-
ciated proteins are identified by MS. Two large-scale studies in yeast
have been published (Gavin et al. 2002; Ho et al. 2002). In both
cases a large number of protein-protein interactions were detected,
and although there was by no means a perfect overlap between the
two studies and the data already in the public domain, the proteo-
mics strategy appeared to give more complete and reproducible data
than the large-scale Y2H study (Grunenfelder and Winzeler 2002).
To date over 15,000 interactions between almost 5,000 proteins have
been identified in yeast using either the genome wide Y2H or pro-
teomics approaches or by researchers working on individual proteins
of interest. Complex mathematical models are required to pull all of
this data together, and it would appear that the current data set is
not saturated and that combining data from a variety of sources en-
hances and strengthens the quality of the information that can be
generated (Bader and Hogue 2002).
As an alternative to using tagging and immunoprecipitation strate-
gies to identify protein complexes, there have been several studies
where sub-cellular fractionation has been used. These studies seek to
define protein function by the identification of which organelle they
are found in, and a number of studies have analysed the protein
composition of the nucleolus (Andersen et al. 2002), the phagosome
40 R. Solari
(Gagnon et al. 2002) and the Golgi (Bell et al. 2001). We are start-
ing to redefine the anatomy of the cell by the molecular composition
of its constituent parts.
Generating models of protein-protein interactions is highly com-
plex, but it is an early step. Linking the protein clusters to known
biochemical pathways or cellular structures is now possible, and one
needs to integrate data from proteomics with genetics, biochemistry
and physiology. This takes protein annotation to new levels of so-
phistication in trying to build the links between the genome and
function. However, proteins networks are not static, they change dy-
namically with changing conditions within the cell and we need to
overlay on this basic framework a set of kinetic information. We are
still far from such complex models.
2.5.7 Metabolomics
The central dogma teaches us the DNA makes mRNA that makes
proteins. Unfortunately, whereas the dogma stops there, biology does
not. Proteins control biological events in the cell and many of these
biological events depend upon the control of metabolites within the
cell. The newest field of "-omics" is metabolomics, which is the
comprehensive measurement of metabolites in biological fluids and
tissues using nuclear magnetic resonance (NMR) and MS tech-
niques. If one truly hopes to understand how genes function in cells
or tissues then one must begin to integrate genomics, proteomics
and metabolomics (Nicholson et al. 2002). Whereas changes in gene
or protein expression inform us as what may be happening in re-
sponse to a biological stimulus, measuring changes in metabolite
profiles tells us what has actually happened.
2.5.8 Functional Analysis
All of the techniques mentioned so far attempt to identify the bio-

logical function of a gene by predictive bioinformatics, or changes
in gene expression or protein expression following a particular stim-
ulus and trying to correlate these observations with the biological
outcome. However, there are alternatives which involve defining a

gene function by directly measuring its influence on a biological
mechanism. Such models may involve setting up a biological assay
then testing libraries of genes for their ability to modify the biologi-
cal response in the assay. For example, one may wish to identify
genes that play a role in a particular signal transduction pathway.
One example might be a cellular system where the ligand for a cell
surface receptor activates a pathway that ultimately turns on a repor-
ter gene that can be quantified. Libraries of cDNAs can be intro-
duced into these cells, either by transfection of plasmid DNA or by
using viral vectors such as adenovirus, and clones that modify the
pathway can thus be identified. Such strategies are called "expres-
sion cloning" and have been successfully used to identify many on-
cogenes, growth factors and their receptors as well as components of
signalling pathways.
2.6 Target Validation Technologies
There are many avenues of research that might identify a potential

drug target, but how many of these potential targets will be worth
starting a drug discovery programme on? One would need to have
some evidence that the "target" is implicated in disease pathology or
that activation or inhibition of the target would modify disease
pathology or provide symptomatic relief. Finally one would hope to
get some idea of potential safety or toxicity issues. Addressing these
issues has come to be known as "target validation" and there is a
wide range of techniques at the disposal of the drug discovery scien-
tist. The target discovery strategies described above should be de-
signed to provide a hypothesis, and it is the role of target validation
studies to test such hypotheses in experimental models.
If the potential target has been discovered by mRNA expression
profiling studies, the first validation exercise is the confirmation that
the mRNA is truly overexpressed in tissues of interest by real time
(RT) PCR, Northern blots and in situ hybridisation (ISH). Following
this confirmation one would want to be sure that the mRNA expres-
sion correlates with protein overexpression by immunohistochemis-
try (IHC). In cancer therapeutics, selective overexpression of a cell
42 R. Solari
surface protein in a tumour may be enough to design a targeting

antibody that delivers a radionuclide or cytotoxic drug to the tu-
mour. The biological function of the target may not be important in
this case. However, for most therapeutics one needs to confirm that
the target has some role in the pathology or that it might be useful
for modifying symptoms of the disease.
In the case that the genomics, bioinformatics or proteomics plat-
forms identify a potentially novel target of unknown function, the
cDNA is usually cloned and the protein product tested in appropriate
cellular or whole animal models for biological activity. On the other
hand, the function of the gene product may already be known or de-
duced from the literature. This first step "validates" that the novel
target has a defined biological activity that may be consistent with a
role in pathology.
The second step is to confirm that this new protein has a role in a
disease pathology or that expressing or inhibiting the protein may
have some beneficial effect on the disease. This second step of vali-
dation requires disease models and tools or reagents to test the hy-
pothesis. Perhaps the oldest target validation tools are antibodies.
Specific antibodies that block the activity of cell surface and se-
creted proteins can be quite difficult to make but are valuable re-
agents and more and more frequently are becoming therapeutic mol-
ecules in their own right.
An alternative strategy is to reduce target protein expression by
blocking or knocking down the level its mRNA. Three major tech-
nologies have emerged for mRNA knock-down, antisense oligonu-
cleotides, ribozymes and siRNA (Opalinska and Gewirtz 2002).
These three types of reagents work differently, although the end re-
sult is similar. Antisense are short oligonucleotides complementary
to the specific mRNA sequence that one wants to eliminate. These
antisense oligonucleotides hybridise with the mRNA, and the dou-
ble-stranded region so created either blocks translation or targets the
mRNA for degradation via an RNase enzyme. Ribozymes are also
single stranded oligonucleotides that bind to specific mRNAs by
complementary base pairing. However, the ribozyme has a non-com-
plementary loop that forms a hairpin or hammerhead structure with
catalytic activity capable of cleaving the mRNA target. The most re-
cent development in this field has been RNA interference. This is
the process whereby small double-stranded (ds )RNA molecules (21-

15 base pairs) target mRNA for destruction in a sequence-specific
fashion. These short dsRNAs are known as short interfering RNA or
siRNA. RNA interference appears to be a natural host defence pro-
cess and starts with the processing of long (500-1,000 base pair)
dsRNA into short 21-25 base pair fragments by an enzyme called
DICER. When these short fragments become incorporated into the
large nuclease complex called the RNA-induced silencing complex
(RISC), they unwind, hybridise to their complementary mRNA and
catalyse the sequence-specific destruction. Long double-stranded
RNAs have been used for some time to silence the expression of tar-
get genes in a variety of organisms and cell types (e.g. worms, fruit
flies, and plants). In mammalian cells, however, introduction of long
dsRNA (>30 base pairs) initiates a potent antiviral response, exem-
plified by non-specific inhibition of protein synthesis and RNA de-
gradation. The mammalian antiviral response can be bypassed, how-
ever, by the introduction or expression of siRNAs. Although RNA
knock-down strategies are relatively effective in vitro and there are a
variety of cellular delivery techniques, so far they have proved diffi-
cult to use in in vivo models, which limits their value as target vali-
dation tools.
In terms of gene knock-out technology for target validation, the
gold standard is still the mouse. The ease with which genes can be
manipulated in the mouse and its convenience as an experimental
model make it the ideal choice. Increasingly the mouse technology
is becoming more sophisticated and gene knock-outs and overex-
pression in specific tissues and at specific times are now possible.
Coupling this to disease challenge models can provide a sophisti-
cated framework for testing the therapeutic validity of a new target.
However, the mouse is by no means perfect, and the ultimate tar-
get validation tool for a drug target is a drug itself in an appropriate
model of disease. The availability of high-throughput screening and
large chemical libraries has made it possible to identify small mole-
cule agonists or antagonists of potential targets with unprecedented
speed, and the hope is that it may be possible to validate the target
using these chemical entities. However, the reality is that "hits"
from chemical screening rarely have sufficient potency, selectivity or
the appropriate physico-chemical properties to make them suitable
44 R. Solari
for testing in animal models. A great deal of chemical optimisation

needs to be done before the small molecule can be used to test a
biological hypothesis with any expectation of a sensible outcome.
The "Catch-22" is that chemists are usually not motivated to per-
form long and difficult lead optimisation work on a target that is
poorly validated in the first place.
2.7 Conclusions
Potential new drug targets can come form a wide range of sources.
Traditional, biology-driven hypothesis testing research in the public
domain is still the greatest source of such targets. Increasingly, we
are seeing the emergence of genome-wide or large-scale approaches
to studying biology that are largely hypothesis free. These huge data
collecting exercises have the potential to make profound ground-
breaking observations forcing our understanding of biology forward.
However, all too often, poor experimental design, questionable math-
ematical analysis and lack of reproducibility and comparability be-
tween the various technology platforms are holding back progress.
Target validation relies heavily on skills in molecular cell biology,
physiology and pharmacology and there are few shortcuts to under-
standing complex biological systems. Nevertheless, we now have at
our disposal an impressive toolbox of techniques that allow us to
probe biological systems as never before.
References
Andersen JS, Lyon CE, Fox AH, Leung AK, Lam YW, Steen H, Mann M,
Lamond AI (2002) Directed proteomic analysis of the human nucleolus.
CUff BioI 12:1-11
Bader GD, Hogue WV (2002) Analysing yeast protein-protein interaction
data obtained from different sources. Nat Biotechnol 20:991-997
Bell AW, Ward MA, Blackstock WP, Freeman HN, Choudhary JS, Lewis
AP, Chotai D, Fazel A, Gushue IN, Paiement J, Paley S, Chevet E, Lafre-
niere-Roula M, Solari R, Thomas DY, Rowley A, Bergeron JJ (2001)
Proteomics characterization of abundant golgi membrane proteins. J BioI
Chern 276:5152-5165
Bhattacharjee A, Richards WG, Staunton J, Li C, Monti S, Vasa P, Ladd C,

Beheshti J, Bueno R, Gillette M, Loda M, Weber G, Mark EJ, Lander
ES, Wong W, Johnson BE, Golub TR, Sugarbaker DJ, Meyerson M
(2001) Classification of human lung carcinomas by mRNA expression
profiling reveals distinct adenocarcinoma subclasses. Proc Nat! Acad Sci
USA 98:13790-13795
Chandrasekharan NV, Dai H, Roos KL, Evanson NK, Tomsik J, Elton TS,
Simmons DL (2002) A cyclooxygenase-I variant inhibited by acetamino-
phen and other analgesic/antipyretic drugs: cloning, structure and expres-
sion. Proc Natl Acad Sci USA 99: 13926-13931
De Strooper B, Konig G (1999) Alzheimer's disease: a firm base for drug
development. Nature 402:471-472
DiMasi JA (200 I) New drug development in the United States from 1963 to
1999. Clin Pharmacol Ther 69:286-296
Dixon RA, Kobilka BK, Strader DJ, Benovic JL, Dohlman HG, Frielle T,
Bolanowski MA, Bennett CD, Rands E, Diehl RE et al (1986) Cloning
of the gene and the cDNA for mammalian p-adrenergic receptor and
homology with rhodopsin. Nature 321:75-79
Drews J (1996) Genomic sciences and the medicine of tomorrow. Nat Bio-
technoI14:1516-1518
Drews J, Ryser S (1997) Classic drug targets. Nat Biotechnol 15:1318-1319
Gagnon E, Duclos S, Rondeau C, Chevet E, Cameron PH, Steele-Mortimer
0, Paiement J, Bergeron n, Desjardins M (2002) Endoplasmic reticulum-
mediated phagocytosis is a mechanism of entry into macrophages. Cell
110:119-131
Gavin AC, Bosche M, Krause R, Grandi P, Marzioch M, Bauer A, Schultz J,
Rick JM, Michon AM, Cruciat CM, Remor M, Hofert C, Scheider M,
Brajenovic M, Ruffner H, Merino A, Klein K, Hudak M, Dickson D,
Rudi T, Gnau V, Bauch A, Bastuck S, Huhse B, Leutwein C, Heurtier
MA, Copley RR, Edelmann A, Querfurth E, Rybin V, Drewes G, Raida
M, Bouwmeester T, Bork P, Seraphin B, Kuster B, Neubauer G, Superti-
Furga G (2002) Functional organization of the yeast proteome by system-
atic analysis of protein complexes. Nature 415: 141-147
G6nczy P, Echeverri C, Oegema K, Coulson A, Jones SJ, Copley RR, Du-
peron J, Oegema J, Brehm M, Cassin E, Hannak E, Kirkham M, Pichler
S, Flohrs K, Goessen A, Leidel S, Alleaume AM, Martin C, Ozlu N,
Bork P, Hyman AA (2000) Functional genomics analysis of cell division
in C. elegans using RNAi of genes on chromosome III. Nature 408:331-
336
Grunenfelder B, Winzeler E (2002) Treasures and traps in genome-wide
data sets: case examples from yeast. Nat Rev Genet 3:653-661
Ho Y, Gruhler A, Heilbut A, Bader GD, Moore L, Adams SL, Millar A,
Taylor P, Bennett K, Boutilier K, Yang L, Wolting C, Donaldson I,
Schandorff S, Shewnarane J, Vo M, Taggart J, Goudreault M, Muskat B,
Alfarano C, Dewar D, Lin Z, Michalickova K, Willems AR, Sassi H,
46 R. Solari: Target Discovery and Validation
Nielsen PA, Rasmussen KJ, Andersen JR, Johansen LE, Hansen LH, Jes-
persen H, Podtelejnikov A, Nielsen E, Crawford J, Poulsen V, Sorensen
BD, Matthiesen J, Hendrickson RC, Gleeson F, Pawson T, Moran MF,
Durocher D, Mann M, Hogue CW, Figeys D, Tyers M (2002) Systematic
identification of protein complexes in Saccharomyces cerevisiae by mass
spectrometry. Nature 415: 180--183
Hopkins AL, Groom CR (2002) The druggable genome. Nat Rev Drug Dis-
cov 1:727-730
International Human Genome Sequencing Consortium (2001) Initial se-
quencing and analysis of the human genome. Nature 409:860-922
Ito T, Chiba T, Ozawa R, Yoshida M, Hattori M, Sakaki Y (2001) A com-
prehensive two-hybrid analysis to explore the yeast protein interactome.
Proc Nat! Acad Sci USA 98:4569--4574
Kujubu DA, Fletcher BS, Varnum BC, Lim RW, Herschman HR (1991)
TIS 10, a phorbol ester tumour promoter-inducible mRNA from Swiss
3T3 cells, encodes a novel prostaglandin synthase/cyclooxygenase homo-
logue. J Bioi Chern 266:12866-12872
Nicholson JK, Connelly J, Lindon JC, Holmes E (2002) Metabonomics: a
platform for studying drug toxicity and gene function. Nat Rev Drug Dis-
cov 1:153-162
Opalinska JB, Gewirtz AM (2002) Nucleic acid therapeutics: basic princi-
ples and recent applications. Nat Rev Drug Discov 1:503-514
Uetz P, Giot L, Cagney G, Mansfield TA, Judson RS, Knight JR, Lockshon
D, Narayan V, Srinivasan M, Pochart P, Qureshi-Emili A, Li Y, Godwin
B, Conover D, Kalbfleisch T, Vijayadamodar G, Yang M, Johnston M,
Fields S, Rothberg JM (2000) A comprehensive analysis of protein-pro-
tein interactions in Saccharomyces cerevisiae. Nature 403:623-627
Venter JC, Adams MD, Myers EW, Smith HO et al (2001) The sequence of
the human genome. Science 291: 1304--1351
Xie WL, Chipman JG, Robertson DL, Erikson RL, Simmons DL (1991) Ex-
pression of a mitogen-responsive gene encoding a prostaglandin synthase
is regulated by alternative splicing. Proc Nat! Acad Sci USA 88:2692-
2696
Zambrowicz BP, Sands AT (2003) Knockouts model the 100 best-selling
drugs - will they model the next 100? Nat Rev Drug Discov 2:38-51
Zhang Y, Proenca R, Maffei M, Barone M, Leopold L (1994) Positional
cloning of the mouse obese gene and its human homologue. Nature
372:425--432
Zhu H, Bilgin M, Bangham R, Hall D, Casamayor A, Bertone P, Lan N,
Jansen R, Bidlingmaier S, Houfek T, Mitchell T, Miller P, Dean RA, Ger-
stein M, Snyder M (2001) Global analysis of protein activities using pro-
teome chips. Science 293:2101-2105
3 Noninvasive Imaging
in Drug Discovery and Development
M. Rudin, P. Allegrini, N. Beckmann, H.-U. Gremlich,

R. Kneuer, D. Laurent, M. Rausch, M. Stoeckli
3.1 Introduction................................. 47
3.2 Upstream Drug Discovery: Target Validation . . . . . . . . . . . . 50
3.3 Characterization of the Disease Phenotype. . . . . . . . . . . . . . 55
3.4 Assessment of Drug Efficacy . . . . . . . . . . . . . . . . . . . . . . 58
3.5 Drug Biodistribution and Pharmacokinetics . . . . . . . . . . . . . 61
3.6 Imaging Biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.7 PotentiallLimitations of Imaging Approaches . . . . . . . . . . . . 64
3.8 Conclusion.................................. 67
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.1 Introduction
Traditionally, medical imaging is applied to radiological diagnostics,

i.e., to characterize a disease phenotype based on morphological or
physiological readouts. Criteria determining image quality and,
hence, its diagnostic value are spatial resolution and contrast-to-
noise ratio (CNR), the ability to discriminate an anatomical structure
from its environment. Image contrast is based on the biophysical
properties of tissue such as absorption of radiation in X-ray, or pro-
ton density and spin relaxation in magnetic resonance imaging
(MRI). Structure definition can be enhanced by administration of ex-
ogenous contrast agents, which are electron-dense materials for X-
ray and para- or superparamagnetic compounds for MRI, the con-
48 M. Rudin et al.
trast enhancement being governed by both the properties of the con-

trast agent and the tissue microstructure. The dynamic measurement
of contrast changes induced by administration of a contrast agent
provides information on physiological tissue parameters such as tis-
sue perfusion, vascular permeability, and function of excretory or-
gans. When combining the contrast-enhancing principle (reporter
group) with a target-specific carrier moiety (receptor ligand, anti-
body, cell), specific information on target expression and function
can be obtained. Such target-specific or molecular imaging ap-
proaches have raised considerable interest both from a diagnostic
and therapeutic point-of-view (Rudin and Weissleder 2003). Consid-
ering the broad spectrum of applications of modem imaging technol-
ogies their increasing impact on drug discovery and development
programs is not surprising.
Today, the number of drug targets (receptors, enzymes), for which
drugs exist or which are currently being worked on is of the order
of a few hundreds. Due to the impact of functional genomics, this
number is to increase significantly prompting the need to develop ef-
Geoomics_
Model Target
Q
systems Selection &
validation
Relevance of Readouts on
Clinical drug target: - PK
pathology/ - expression levels - Efficacy
patho- (up-/down -regulation) - Safety
physiology - function - Mechanism
Biomarkers
Fig. 1. Imaging in drug discovery and development. The recent develop-

ments in target-specific (molecular) imaging has provided a toolset for target
validation studies. The downstream phases (lead optimization and beyond)
are predominantly covered by structural and physiological imaging ap-
proaches providing information on the disease phenotype as well as drug ef-
ficacy and safety. The identification and validation of biomarkers to be used
for clinical drug evaluation constitutes is an area of intense research. In fu-
ture, molecular imaging approaches will have an increasing impact on all
phases of the process
Noninvasive Imaging in Drug Discovery and Development 49
ficient filters for the selection of the most promising targets with re-
gard to disease relevance and "drugability". While noninvasive imag-
ing is unlikely to playa major role in this target selection process, it
will become an important tool for target validation providing tem-
porospatial information on target expression levels and/or function in
relation to a disease process (Fig. 1). Target-specific methods de-
signed to study such molecular events in an intact organism are cur-
rently being developed at a rapid pace (Rudin and Weissleder 2003).
The large majority of imaging applications to date are focused on
the evaluation of lead compounds in animal models of human dis-
ease (Fig. 1). Both qualitative and quantitative morphological, physi-
ological, and metabolic imaging methods have been extensively used
to characterize pathogenesis (Rudin et al. 1999; Beckmann et al.
2000, 2001; Rudin and Weissleder 2003) and to evaluate the efficacy
of a therapeutic intervention. Molecular imaging approaches will
provide complementary readouts. Obviously, analogous methods can
be used to assess drug safety. The noninvasiveness of the imaging
approach offers a number of distinct advantages: (1) The structural
and physiological data can be correlated to clinical readouts, (2)
chronic disease process can be monitored in the same individual en-
hancing the statistical relevance of the information gathered, (3)
baseline measurements can be carried out prior to drug evaluation,
allowing stratification of treatment groups, and (4) similar methods
can be applied to clinical drug evaluation. The last point holds for
clinically established imaging modalities such as computed tomogra-
phy (CT), ultrasound, MRI and nuclear imaging methods, but not
for optical imaging technologies, the use of which is currently lim-
ited to animal studies.
Clinical development of drug candidates is time consuming and
costly. Thus, before entering a full-blown development program, in
particular when targeting chronic diseases with late clinical end-
points, it is important to identify reliable early biomarkers of drug
efficacy. There is intensive research going on in trying to define po-
tential biomarkers for drug efficacy and safety involving readouts at
the level of gene transcription, gene translation or regarding drug-in-
duced metabolic alterations (Nicholson et al. 2002), all requiring tis-
sue and/or body-fluid sampling. In this context, biomarkers based on
noninvasive imaging readouts are highly attractive.
50 M. Rudin et al.
3.2 Upstream Drug Discovery: Target Validation
The possibility to image molecular events under in vivo conditions

lends itself for validation of a potential drug target. The obvious
approach would be to develop a labeled reporter ligand that directly
interacts with the target with high selectivity (Fig. 2). The develop-
ment of such probes involves steps analogous to drug development:
the specificity of the interaction, the biodistribution and pharmacoki-
netics, and the sensitivity for detection have to be characterized.
This approach is laborious and time consuming and probably only
feasible for selected high-value targets. Alternatively, a generalized
reporter system could be developed that might be used to visualize
many different biological processes using the same reporter probe
and different pretargeting molecules (Massoud and Gambhir 2003).
This indirect approach has been used to image regulation of gene
expression (Louie et al. 2000; Weissleder et al. 2000; Carlsen et al.
2002; Green et al. 2002; Ciana et al. 2003), protein-protein interac-
Drug Target Pathway System response
c)
a)
~ Structural phenotype
~---+ f) ---+ Physiological/functional

phenotype
~ Metabolic phenotype
Fig. 2. Imaging targets: labeling of the drug (a) allows the study of its bio-
distribution, pharmacokinetics, and the assessment of receptor occupancy.
Target expression is monitored using target-specific reporter probe (b) that
bind to the active or to an allosteric site of the drug target (A: receptor, en-
zyme). Alternatively, target expression can be studied using reporter genes
under the control of the promoter of the respective target (Prom A; c). Re-
ceptor function leads to altered pathway fluxes (d) and may induce gene ex-
pression that can be monitored using reporter genes (e). Finally, the re-
sponse to the drug-target interaction is translated into microscopic and
macroscopic equivalents (j), i.e., a structural, physiological, and/or metabolic
phenotype
tions (Luker et al. 2002; Ray et al. 2002), or cell trafficking (Mandl
et al. 2002; Scheffold et al. 2002). A disadvantage of the method is
that foreign reporter genes have to be introduced into the cell, which
might constitute a limitation in intact organisms, as the gene prod-
ucts might affect the cell function.
Direct target visualization using labeled reporter ligands is the
classical molecular imaging application. Nuclear imaging and in par-
ticular positron emission tomography (PET) have been used for
years to study receptor occupancy and, hence, expression using radio-
labeled reporter systems. PET ligands, isotopically labeled with the
positron emitting nuclei 18F or 11 C, have been developed for most of
the neurotransmitter systems, offering an attractive toolset to the
neuropharmacologist (Langstrom et al. 1999). With the development
of optical imaging methods such as near-infrared fluorescence imag-
ing (NIRF) (Weissleder et al. 1999), target-specific probes have been
developed using fluorescent dyes as reporter moiety. For example,
by labeling the long-lived somatostatin analog octreotide with a cya-
nine dye, expression of somatostatin receptor in murine tumor xeno-
grafts could be demonstrated (Becker et al. 2001). Similar experi-
ments have been carried out using 111 In labeling and single photon
emission computer tomography (SPECT) (van Royen et al. 1996).
Alternatively, the reporter moiety has been linked to an antibody or
antibody fragment such as to phage-derived human antibody frag-
ments with a high affinity for the extra-domain B (ED-B) of fibro-
nectin, a marker of angiogenesis (Birchler et al. 1999). A critical as-
pect of these direct labeling strategies is the discrimination of sig-
nals of specific and unspecific ligand binding, requiring (1) high
specificity and affinity of the ligand and (2) rapid elimination of the
unspecifically bound fraction. Alternative approaches have been de-
veloped using probes that are selectively activated by the target (re-
ceptor, enzyme) such as the protease probes developed for NIRF
imaging (Mahmood et al. 1999; Weissleder et al. 1999; Tung et al.
2000, 2002). Protease-specific peptide sequences with Cy5.5 groups
attached are linked to a biocompatible macromolecular backbone.
Fluorescence is quenched unless the peptide sequence is cleaved by
the protease, leading to an amplification of the fluorescence signal
by two orders of magnitude or even more. This allows direct assess-
ment of the enzyme activity under in vivo conditions as shown for a
52 M. Rudin et al.
Fig. 3. Near infrared fluorescence imaging of protease activity in prostate

carcinoma orthotopically implanted in a murine model. In the absence of
protease activity, the fluorescent dyes are bound to a macromolecular back-
bone and fluorescence is quenched due to the proximity of the dye groups.
In the presence of the protease, the specific peptide sequences are cleaved,
releasing monomeric dye molecules leading to an increase of fluorescence
intensity by more than two orders of magnitude. Protease ligands were origi-
nally developed by Weissleder et al. (1999). Numbers in the figure indicate
tumor volume in mm 3
matrix metalloproteinase 2 (MMP2)-specific probe in a murine tu-

mor model (Fig. 3).
Although such direct imaging approaches yield highly specific in-
formation on the expression of a drug target, their applicability is
limited by just this specificity - each target requiring its own specif-
ic probe. It is therefore not surprising that indirect imaging methods
using reporter genes (Fig. 2), which have been so successfully ap-
plied in in vitro assays, are a focus area of molecular imaging devel-
opments. Reporter gene imaging is used as a tool to study the regu-
lation and the temporospatial profile of gene expression. Well-char-
acterized reporter gene/reporter probe systems used for in vivo imag-
ing are listed in Table 1. Reporter gene imaging is used for monitor-
ing the expression of endogenous genes or transgenes, for gene
marking of cells, and to image molecular interactions, e.g., protein-
protein interactions, as illustrated with some examples.
The modulated expression of endogenous albumin in transgenic
mice has been monitored using herpes simplex virus 1 tyrosine kinase
(HSV I-tk) under control of the albumin promoter in combination with
the PET ligand 18F-fluoro-penciclovir (Green et al. 2002). Biolumines-
cence generated by the interaction of firefly luciferase (Flue) with D-
luciferin is a frequently used reporter system in transgenic animal re-
search. Examples comprise the monitoring of nuclear factor (NF)-KB
activity using Flue under the control of the NF-KB promoter (Carlsen
et al. 2002) and the study of the dynamics of estrogen receptor activity
Table 1. Reporter gene systems developed for in vivo imaging
Reporter gene product Reporter ligand Imaging Reference

modality
Herpes simplex virus I 131 I labeled FIAU SPECT Tjuvajev

thymidine kinase et al. 1996
(HSVI-tk) 18F_fluoro-pencicIovir PET Iyer et al.
2001
18F-fluoro-guanine PET Namavari
derivatives et al. 2000
Dopamine type 2 18F_ fluoroethylspiperone PET MacLaren
receptor (D2R) (FESP) et al. 1999
ji'-Galactosidase EgadMe MRI Louie
(Pyranosy I-capped et al. 2002
Gd-complex)
Mutated transferring Monocrystalline iron MRI Weissleder
receptor oxide nanoparticles et al. 2000
(MION)
Firefly luciferase D-Luciferin Biolumi- Contag
nescence et al. 1997
Green fluorescent Biolumi- Benaron
protein (GFP) nescence et al. 1997
EgadMe, (1-(2-(ji'-galactopyranosyloxy)propyl)-4,7, 10-tris(carboxymethyl)-

I, 4,7,IO-tetraazacyclododecane) gadolinium(III); FIAU, 2'-fluoro'-2'-deoxy-
I-ji'-D-arabinofuranosy1-5-iodo-uracil.
54 M. Rudin et al.
and its dependence on 17p-estradiol (Ciana et al. 2003). Correspond-

ing experiments have been carried out using contrast-enhanced MRI.
The assessment of p-galactosidase (p-gal) activity under the control
of the Rous sarcoma virus promoter has been studied in Xenopus lar-
vae using pyranosyl capped gadolinium (Gd) chelates, which undergo
a change in T 1 relaxivity after cleavage of the sugar residue by p-gal
(Louie et al. 2000). A different reporter system has been used to moni-
tor gene expression in murine tumor models: the expression of a mu-
tated ferritin receptor led to massive accumulation of monocrystalline
iron oxide nanoparticles (MION), a superparamagnetic reporter moi-
ety, in the respective cells (Weissleder et al. 2000). Similar reporter
gene approaches have been developed to study molecular interactions
such as protein-protein interactions (Luker et al. 2002; Ray et al.
2002). These examples clearly illustrate that both direct and indirect
target imaging provide direct quantitative information on target ex-
pression levels in vivo. It is beyond doubt that such methods will be
extensively used in pharmacological research.
Direct visualization of target expression is complemented by
imaging of target function, i.e., of downstream processes induced by
target activation (Fig. 2). As an example we discuss apoptosis. Sev-
eral approaches have been reported to assess apoptosis, all of them
targeting aminophospholipids, such as phosphatidylserine (PtdS),
which are redistributed from the inner to the outer leaflet of the cell
membrane during the apoptotic process. Potential PtdS targeting
moieties are annexin V and the C 2 domain of synaptotagmin I. Es-
sentially all modalities suited for molecular imaging have been ap-
plied to visualize apoptosis: SPECT with 99ffiTc-Iabeled annexin V
(Blankenberg et al. 1998), MRI in combination with iron-oxide na-
noparticles coupled to the C 2 domain of synaptotagmin I (Zhao et
al. 2001) or to annexin V (Schellenberger et al. 2002), and NIRF
imaging with CyS.S.-Iabeled annexin V (Petrovsky et al. 2003). An
indirect labeling approach uses Fluc as reporter gene that is acti-
vated by apoptosis-induced caspase-3 specific activity (Laxman et
al. 2002). A fusion protein has been designed, in which estrogen re-
ceptor regulatory domains (ER) are coupled to both the C- and N-
terminus of Flue, thereby suppressing bioluminescence activity. By
introducing a protease cleavage site between Fluc and ER, the repor-
ter is activated upon expression of the caspase.
3.3 Characterization of the Disease Phenotype
Imaging of a disease phenotype in animal models is the direct corre-

late to clinical diagnostic imaging. Today, MRI has evolved to the
standard tool for morphological imaging in small animals due to the
high spatial resolution provided (typically pixel dimensions of the
order of 100 11m) and due to its excellent soft tissue contrast charac-
teristics (Rudin et al. 1999). For specific applications, complemen-
tary modalities might be superior, e.g., X-ray CT for the quantitative
assessment of bone density and architecture.
There is a vast literature on using imaging and in particular MRI for
pathomorphological characterization of animal models of human dis-
ease. Therefore, we will focus on one example only, neurodegenerative
disorders. Animal models of neurodegenerative disorders typically in-
volve focal cerebral lesions (Fuxe and Ungerstedt 1976; Oh et al. 1994).
More recently a significant number of models involving genetically
engineered mice have been developed such as mice over-expressing
human amyloid precursor protein (Sturchler-Pierrat et al. 1997) as mod-
el of Alzheimer's disease (AD). The APP23 mouse model reproduces
neuropathological changes associated with AD such as high levels of
amyloid plaques with a core of /l-amyloid (Ap). These deposits start
to appear with 6 months of age predominantly in the neocortex and
hippocampus and increase in number and size until the age of
24 months when they occupy substantial portions of these brain struc-
tures (Sturchler-Pierrat et al. 1997; Price et al. 1998). Plaque diameters
are of the order of 100 11m and have been detected in human brain spec-
imen using high resolution anatomical MRI (Benveniste et al. 1999).
Nevertheless, this MR microscopic approach is not easily transferable
to the in vivo situation due to prohibitive measurement times and prob-
ably altered contrast characteristics in vivo. Alternative MRI ap-
proaches to visualize plaques use plaque-specific contrast agents such
as labeled AfJ peptides (Wadghiri et al. 2002) or a gadolinium-labeled
AfJ peptide-putrescine complex (Poduslo et al. 2002) to enhance the
CNR. Nevertheless, these approaches suffer from significant limita-
tions (low sensitivity of MRI, low blood-brain barrier penetration)
and the use of plaque-specific MRI contrast agents in the clinics is cur-
rently not possible. Small molecular PET ligands are certainly more
promising in this regard (Agdeppa et al. 2001; Shoghi-Jadid et al. 2002).
56 M. Rudin et al.
MRI has been used both clinically and preclinically to provide

secondary structural readouts such as measures of brain atrophy (to-
tal brain volume, volume of defined brain structures such as hippo-
campus, ventricular volumes) or of indicators of microstructural
changes of brain parenchyma such as the apparent water diffusion
coefficient, which is influenced by the ratio of extracellular versus
intracellular volumes (Benveniste et al. 1992). Nevertheless, at least
for the APP23 model of AD, these structural changes seem to reveal
only late-stage pathology and hence are of limited value for disease
phenotyping.
The clinical hallmark of AD is cognitive impairment. Progressive
deterioration of behavioral/cognitive functions has also been re-
ported for aged APP23 transgenic mice (Lalonde et al. 2002) sug-
gesting a link to the pathomorphological and pathophysiological
changes. Such functional deficits can be quantitatively assessed
using functional MRI (fMRI) by analyzing the response to standard-
ized pharmacological or peripheral sensory stimuli in transgenic
mice (Mueggler et al. 2002 a, b). fMRI measures the hemodynamic
consequences such as changes in local cerebral blood volume (CBV)
or blood oxygenation levels (BOLD) associated with brain activity.
The CBV response to three different stimuli was significantly re-
duced in 15 and 24 months old APP23 transgenic animals as com-
pared to age-matched littermates. Experiments using a vasodilatory
challenge revealed that the reduced response is most likely due to a
compromised vascular reserve capacity caused by severe cerebral
amyloid angiopathy. These results demonstrated that functional phe-
notyping complements more conventional morphological character-
ization and, at least for APP23 mice, displayed a high sensitivity to
pathophysiological alterations.
Many pathological events involve cell migration. Examples are
the infiltration of monocytes and lymphocytes to a site of inflamma-
tion or metastasis formation in tumor models. Several imaging ap-
proaches that involve cell-specific labeling strategies have been de-
veloped or are currently under development. Using ultra-small iron
oxide nanoparticles (USPIO), which are internalized by phagocytotic
macrophages, the migration of these monocytes to sites of inflamma-
tion could be visualized by MRI in models of experimental autoim-
mune encephalomyelitis (EAE) (Dousset et al. 1999; Rausch et al.
Fig. 4. Macrophage labeling in model of experimental autoimmune encepha-

lomyelitis (EAE) in rats recorded during the acute phase of brain inflamma-
tion. After systemic administration, ultra-small particles of iron-oxide (US-
PIO; Sinerem, Laboratoire Guerbet, France) are predominantly cleared by
cells of the monocytes phagocytotic system. This allows the monitoring of
macrophage migration to the site of inflammation. Transverse sections at the
level of cerebellum and brain stem of an EAE rat were recorded at day 11
post immunization with myelin basic protein. Focal dark areas were ob-
served in the brainstem 24 h after i.v. administration of USPIOs (right), re-
flecting significantly increased R2 relaxation due to the accumulation of US-
PIO-Ioaded macrophages. In contrast, the animal receiving no contrast agent
did not display hypointense areas in the brain
2003; Fig. 4), brain ischemia (Rausch et al. 2001, 2002), soft tissue
inflammation (Kaim et al. 2002), and kidney transplantation (Beck-
mann et al. 2003). USPIO labeling was also applied to monitor the
migration of labeled stem cells to the site of ischemic brain damage
(Hoehn et al. 2002) or to track progenitor cells in a model of trau-
matic spinal cord injury (Bulte et al. 1999). An alternative approach
of cell labeling is based on gene marking using reporter genes. Ex-
amples include the visualization of the migration of CD4+ T lympho-
cytes marked with Fluc to the brain of mice with EAE (Costa et al.
2001) or in a model of collagen-induced arthritis in mice (Nakajima
et al. 2001). Expression of Fluc in cancer cells allowed the detection
of micro-metastases in bone marrow (Wetterwald et al. 2002). The
advantage of the bioluminescence approach is the high sensitivity,
its most significant disadvantage is the poor anatomical definition.
58 M. Rudin et al.
Molecular imaging approaches are of increasing importance for the

characterization of a disease phenotype, as they allow the direct visua-
lization of target expression of a target or its up- and down-regulation
in a diseased state. NIRF imaging has been used to demonstrate the
expression of ED-B of fibronectin in murine tumor models (Birchler
et al. 1999). Similarly, the approach has been applied to demonstrate
the expression of various proteases in tumor models (Mahmood et
al. 1999; Weissleder et al. 1999; Tung et al. 2000, 2002) and in mouse
models of atherosclerosis (Chen et al. 2002). Using contrast-enhanced
MRI, the targeting of Gd-diethylene-triamine-pentaacetate (GdDTPA)-
labeled liposomes coupled to intercellular adhesion molecule (ICAM)-
I-specific antibodies to the respective adhesion molecules expressed in
the brain vasculature of EAE mice has been demonstrated (Sipkins et
al. 2000). Similarly, magnetically labeled cationic antibodies coupled
to a small molecular a vP3 ligand highlighted the presence of this cell
surface molecule as a marker of activated endothelial cells in a murine
tumor model (Sipkins et al. 1998).
The complementary used of morphological, physiological, cellu-
lar, and molecular imaging approaches provides a comprehensive
characterization of a disease phenotype, which constitutes the basis
for a robust evaluation of therapy response.
3.4 Assessment of Drug Efficacy
In the past decade, MRI was the most widely used imaging modality
in drug research and therefore most of efficacy data available have
been obtained using this technique. MRI either provides phannaco-
logical endpoints (e.g., the tumor volume following therapeutic inter-
vention) or surrogates, i.e., a measure which correlates with and,
hence, predicts a phannacological endpoint. It is beyond the scope
of this article to review the vast literature on this topic, the reader is
referred to recent review articles (Rudin et al. 1999; Beckmann et al.
2000). Instead we will select a few examples from the CNS and on-
cology area to illustrate the potential of noninvasive imaging for the
assessment of drug efficacy in animal models.
A considerable number of cytoprotective compounds has been
evaluated in focal cerebral ischemia models (for a review see Rudin
et al. 1999). In these studies, drug effects were evaluated based on

infarct volumes as derived from T rweighted images or maps repre-
senting the apparent (water) diffusion coefficient (ADC) , assuming
that a reduced infarct volume translates into an improved clinical
outcome. Alternatively, therapy response was analyzed using hemo-
dynamic (Sauter et al. 1989) or metabolic readouts (Sauter and Ru-
din 1987). A critical piece of information for the evaluation of cyto-
protective therapy is the demonstration of neuronal activity in areas
that have been spared from infarction by drug treatment. Full or par-
tial recovery of function in cytoprotected somatosensory cortices
could in fact be shown in rats following treatment with the calcium
antagonist isradipine (Sauter et al. 2002). However, in 50% of the
animals, brain function in the corresponding area could not be re-
stored within the observation period. Such findings illustrate the im-
portance of functional readouts in the evaluation of anti-ischemic
compounds complementing conventional morphological imaging.
The classical radiological endpoint of anticancer therapy is tumor
volume reduction. Examples of animal studies involving morpho-
metric readouts are the evaluation of treatment of hormone-depen-
dent tumors such as pituitary hyperplasia (Rudin et al. 1988), Dun-
ning prostate tumor (Siegel et al. 1988), or prostate hyperplasia
(Cohen et al. 1995; Haeusler et al. 1996) to name a few. Imaging
obviously offers significant advantages when studying orthotopic tu-
mors, which are difficult to assess otherwise. Nevertheless, tumor
volume is a late readout for drug efficacy, prompting the search for
earlier markers. Changes in gross tumor morphology are preceded
by changes of tumor microstructure, tumor physiology, metabolism,
and at the molecular level, e.g., altered receptor expression profiles
and pathway fluxes. Diffusion-weighted MRI provides information
on changes in cellular volumes (Benveniste et al. 1992). D sing this
technique it could be shown that cytostatic therapy with 1,3-
bis(2chloroethyl)-I-nitrosourea (BCND) transiently increased the ap-
parent water diffusion coefficient in 9L glioma brain tumors (Cheve-
nert et al. 1997) indicative of increased extracellular space due to
apoptotic body formation and cell loss. Similarly, TIp has been pro-
posed as an early MRI indicator of tumor microstructure predicting
treatment efficacy in a study of radiation-induced fibrosarcoma
(RIF)-l tumors treated with cyclophosphamide (Duvvuri et al. 2001).
60 M. Rudin et al.
Tumor proliferation depends on its blood supply, i.e., on the for-

mation of feeding vessels. There are various approaches for assess-
ing tumor angiogenesis: measurement of tumor perfusion, blood vol-
ume, or vascular permeability. Alternatively, markers of activated en-
dothelium such as fibronectin expression can be visualized (Birchler
et al. 1999). Vascular permeability data have been derived from the
measurement of dynamic MRI contrast enhancement following the
administration of an extracellular contrast agent such as GdDTPA.
This approach has been widely used for tumor characterization, e.g.,
for the discrimination between benign and malignant tumors (Knopp
et al. 1999) and for drug efficacy studies (van Dijke et al. 1996;
Brasch et al. 1997). An interesting finding was the apparent discrep-
ancy between effects of the VEGF inhibitor PTK 787 on vascular
permeability and tumor blood volume assessed in a model of renal
carcinoma in rats (Drevs et al. 2002). While the reduced tracer leak-
age indicated anti angiogenic efficacy, an increased tumor blood vol-
ume has been found. Histological analysis using corrosion casts re-
vealed a significant reduction of the number of small vessels
«50 J.lm) while the larger vessel were found to be dilated, account-
ing for both the antiangiogenic efficacy (reduced vessel density) and
the increased blood volume. Hence, a comprehensive in vivo charac-
terization of antiangiogenetic drugs might require additional infor-
mation such as an accurate assessment of the average vessel diame-
ter (Tropres et al. 2001).
Target-specific imaging techniques will be increasingly used to
demonstrate proof-of-concept for individual classes of drugs. As an
example, protease imaging using the already mentioned NIRF li-
gands has been applied to study the efficacy of enzyme inhibitors in
a murine mammary carcinoma model expressing matrix-metallopro-
tease (MMP)-2 (Bremer et al. 2001). Treatment with high doses of
the MMP-2 inhibitor prinomastat led to a significant reduction of
fluorescence activity reflecting reduced MMP-2 activity.
Tools that allow the monitoring of targeted gene delivery are an
important element of gene therapy. This can be achieved by cou-
pling the therapy gene to a targeted reporter moiety. A recently pub-
lished example relates to angiogenesis. The signal transduction path-
way involving Ras-Raf-MEK-ERK seems to play an important role
in neovascularization (Eliceiri and Cheresh 2001). By covalently
binding magnetically labeled antibodies to a small molecular uvlh li-

gand, expression of uv/h, which preferentially occurs in activated en-
dothelial cells, could be monitored in a murine melanoma model (Li
and Bednarski 2002; Li et al. 2002). By coupling the gene for mu-
tated Raf-l, which is unable to bind ATP, to the imaging construct,
significant tumor shrinkage could be induced, demonstrating the effi-
cacy of the gene therapy approach (Hood et al. 2002; Li and Bed-
narski 2002; Li et al. 2002). This study elegantly illustrates the
power of combining a specific diagnostic tool with a targeted carrier
for therapy delivery.
3.5 Drug Biodistribution and Pharmacokinetics
A significant number of drug candidates fail because of inappropri-

ate pharmacokinetic (PK) properties. Hence, early information on
drug biodistribution and PK is essential. Conventionally such data
are obtained from autoradiography studies in rodents requiring isoto-
pically ( 14C, 3 H) labeled compounds. In higher species, such as non-
human primates, and in clinical settings this information is derived
using nuclear imaging methods and in particular PET (Langstrom et
al. 1999; Fischman et al. 2002). With the development of small ani-
mal PET scanners, analogous studies may be carried out in rodents.
The most exciting aspect of PET is the fact that many drugs can be
labeled with IIC or with IsF, i.e., labeling will not or only mini-
mally affect the physicochemical properties of the compound, thus
allowing the monitoring of the drug's biodistribution (Lang strom et
al. 1999; Salazar and Fischman 1999; Fischman et al. 2002). MRI
and optical imaging, on the other hand, use much bulkier reporter
groups that would significantly influence the properties of the target-
specific compound, and are therefore of little value for PK studies.
A basic disadvantage of radiolabeling methods is their low specifici-
ty, i.e., it cannot be distinguished whether the signal originates from
the parent compound or from a metabolite.
Chemical specificity is certainly one of the strengths of magnetic
resonance spectroscopy (MRS), as resonance frequencies are charac-
teristic for the chemical environment of the magnetic nuclei. MRS
has therefore been used by several groups to assess drug PK and me-
62 M. Rudin et al.
tabolism in the intact organism (Griffiths and Glickson 2000; Wolf

et al. 2000). The approach, however, suffers from the inherently low
sensitivity of MRS, which is associated with the low quantum ener-
gies involved in NMR transitions. Consequences are that (1) high
concentrations of analytes are required and (2) spatial resolution is
poor. Drug studies using MRS have been mainly applied to study
the metabolism of fluorinated anticancer (Griffiths and Glickson
2000; Wolf et al. 2000) or psychoactive drugs (Strauss et al. 1997;
Bolo et al. 2000), which are administered at sufficiently high doses
to be detected in vivo by NMR in a sample volume of one to several
milliliters. Fluorinated drugs, in addition, provide the advantage that
there is no interference from a 19F background signal of tissue, thus
optimizing the condition for signal detection. For the large majority
of drugs or drug candidates, the MRS PK studies, however, are not
feasible due to sensitivity constraints.
A final example of monitoring therapy delivery using imaging ap-
proaches is gene therapy. As already discussed, several approaches
are feasible. The therapy gene can be fused to the reporter construct.
It has been shown that the expression of the two genes is highly cor-
related, allowing direct visualization of the expression level of the
therapy gene in tissue (Massoud and Gambhir 2003). Secondly, the
therapy gene construct is coupled to a targeted carrier moiety that
contains a covalently bound reporter group. This approach has been
used for antiangiogenic therapy in tumor models (Hood et al. 2002;
Li and Bednarski 2002; Li et al. 2002). Additional visualization con-
cepts are feasible.
3.6 Imaging Biomarkers
When dealing with chronic diseases, drug studies based on clinical

endpoints are, in general, time consuming. Hence, biomarkers that
give an early indication of drug efficacy or safety would be of high
value. In addition, biomarkers are of relevance when designing and
optimizing personalized therapy and for the selection of patient pop-
ulation for drug trials. In the following we discuss examples from
the areas of CNS disorders and oncology, focusing on clinically es-
tablished imaging modalities such as MRI and PET.
MRI-based biomarkers currently used in the CNS area comprise

structural readouts such as brain atrophy in AD (Smith and Jobst
1996) or depression (Botteron et al. 2002) or functional readouts
such as altered prefrontal cortex functionality in major depression
(Beauregard et al. 1998). PET has been extensively used for charac-
terizing drug efficacy or determining the optimal dosing regimen
using either generalized metabolic readouts such as glucose utiliza-
tion, assessed via the 18F-deoxyglucose method or by studying indi-
vidual neurotransmitter systems (Lang strom et al. 1999). Related in-
formation can be derived by fMRI, which provides valuable infor-
mation whether a drug reaches the brain and prompts a functional
response. This has been demonstrated in animals for a number of
compounds interacting with several neurotransmitter systems: dopa-
minergic, serotonergic, and y-amino-butyrinergic (Chen et al. 1997;
Reese et al. 2000; Xu et al. 2000).
In the oncology area, a number of potential biomarkers to assess
the efficacy of antitumor drugs have been proposed. There is evi-
dence that alteration in the uptake of small molecular extracellular
contrast agents such as GdDTPA by the tumor reflects changes in
microvessel density, tumor blood flow, and leakiness of the vessels,
or a combination thereof (Furman-Haran et al. 1996). Vascular per-
meability measurements have indeed been extensively used to assess
the effects of antiangiogenic VEGF inhibitors both at the preclinical
(Clement et al. 2002) and clinical level (Turetschek et al. 2002). Al-
ternative biomarkers are less mechanism-specific but target general
tumor properties such as metabolism and microstructural changes of
neoplastic tissue. Significant alterations in tumor phospholipids and
energy metabolism induced by drug treatment have been observed
prior to changes in tumor volume (Evelhoch et al. 2000). In line
with this finding, PET revealed significantly reduced glucose utiliza-
tion in gastrointestinal stromal tumors (GIST) already within 12-
24 h after onset of treatment with the tyrosine kinase inhibitor Glee-
vec, indicative of stable therapy response (Van den Abbeele and Ba-
dawi 2002). Finally as already mentioned, the water diffusion coeffi-
cient seems to predict successful tumor therapy: significant increases
of ADC values were observed within a few days of treatment with
the cytostatic drug BCNU, indicative of cell shrinkage, which is at
least in part due to apoptosis (Chevenert et al. 2000).
64 M. Rudin et al.
It is obvious that the cellular and molecular imaging approaches

currently under development will sooner or later translate into clini-
cally applicable biomarkers. However, there are several hurdles to be
overcome. For optical imaging, light scattering properties of tissue is
the ultimate physical limitation, which determines tissue penetration
of light and spatial resolution. In the most favorable case of near-in-
frared light at a wavelength of 730 nm, light attenuation is of the or-
der of 12 dB/cm in brain and 9 dB/cm in skeletal muscle (Weissle-
der and Ntziachristos 2003). In addition, quantitative information on
the local dye concentration requires tomographic techniques which
currently are not available for clinical settings. Another issue when
using molecular imaging techniques in translational research con-
cerns the target-specific probes, which have to be approved for clini-
cal use. This is less problematic for nuclear imaging and in particu-
lar PET approaches requiring only tracer amounts of label, but cer-
tainly holds for MRI and optical imaging probes. Hence, despite the
impressive progress being made in noninvasive small animal molec-
ular imaging, it will take years until this will translate into clinical
useful biomarkers. For the near future, these will still rely and more
classical structural and functional readouts or on the use of radiotra-
cers.
3.7 Potential/Limitations of Imaging Approaches
Today, imaging technologies are established tools in biomedical re-

search and contribute at various stages of drug discovery and devel-
opment. Progress in the last two decades both on the technological
side and with regard to applications has been impressive and is, with
the development of molecular imaging techniques, still going on at a
rapid pace. The ideal imaging technique for pharmaceutical research
should provide high temporospatial resolution, high sensitivity, and
high specificity. In addition, the approach must be noninvasive for
being translatable into the clinical setting. The ideal technology ful-
filling all these criteria does not exist and compromises have to be
searched for.
Spatial resolution is limited by physical or practical limitations of
the specific technology. In MRI, incoherent microscopic motion
(e.g., diffusion processes) during the actual data acquisition (ms) is

the ultimate resolution-limiting factor, i.e., spatial resolution in the
order of micrometers can be achieved, in principle. In fact, MR
images of single cells have been reported (Aguayo et al. 1986). In
practice, the limiting factor, however, is the low sensitivity of the
method, which can only be improved by data averaging resulting in
long measurement times. Typical pixel dimensions are in the order
of 100 11m for animal and I mm for clinical applications for mea-
surement times of seconds or minutes. In PET, the lifetime of the
positron determines positional uncertainty, which, depending on the
radionuclide, is between 0.5 and I mm. Additional errors arise from
deviations from colinearity of the y-rays detected and from technical
limitations in imager design (dimensions of the detector). Resolution
is typically 1.5 mm for animal PET scanners and 3-6 mm for clini-
cal instruments. Finally, for optical imaging (bioluminescence and
NIRF imaging) the positional accuracy is compromised by the light
scattering properties of tissue. While the anatomical definition is
high at the surface, it is severely deteriorated with increasing depth
of the fluorescence or bioluminescence source. In the small animal
tomographic version of NIRF, spatial resolution has been reported to
be of the order of l.5 mm (Ntziachristos et al. 2002; Ntziachristos
and Weissleder 2002). Hence, for structural measurements, MRI is
certainly the method of choice, providing highest anatomical defini-
tion - and in addition excellent soft-tissue contrast. The latter ren-
ders it for many indications superior to X-ray CT (still the gold stan-
dard for anatomical imaging in a clinical setting) and MRI images
are frequently used as anatomical reference image for PET and opti-
cal imaging.
The detection of molecular events occurring with low incidence,
i.e., at low concentration, requires high sensitivity of the in vivo de-
tection method. Here, the nuclear imaging techniques are clearly su-
perior to optical imaging approaches, while MRI is an inherently in-
sensitive technique, which limits its applicability for target-specific
imaging approaches. While significant improvements in detection
sensitivity are conceivable for all approaches, such developments
will not affect the above statement. Sensitivity, i.e., high CNR in
molecular imaging, is provided by the use of target-specific probes,
which should accumulate at the target site while the unspecific back-
66 M. Rudin et al.
ground signal should be minimal. Specific CNR can be improved by

suitable amplification of the label concentration at the target site: ap-
proaches are ligand trapping and ligand activation by the target. In
the former case, the reporter system enters the cell, is modified by
the interaction with the target, and thereafter is unable to leave the
cell. This is, e.g., the case for 18F-fluoro-penciclovir, which is phos-
phorylated by HSVI-tk and thereafter trapped in the cell (lyer et al.
2001). Activatable protease probes represent the second approach:
the respective protease cleaves the peptidic ligand bound to the
macromolecular backbone thereby leading to fluorescence dequench-
ing, which in turn increases the fluorescence intensity by more than
two orders of magnitude (Weissleder et al. 1999). Despite significant
research with regard to optimizing the sensitivity of the reporter
probes per se (improved fluorescence properties, improved magnetic
relaxivity), high signal amplification will be a key to success in mo-
lecular imaging, and improved amplification strategies have to be
developed.
Finally, another enabling factor for molecular imaging is probe
delivery to the target site. This is again less a problem for PET, for
which the probe design consists of isotopic replacement within a
small receptor molecular ligand without affecting its PK properties.
Bioluminescence imaging uses D-Iuciferin as substrate, which can
penetrate cells and may thus be used to study luciferase expression
within the cells. For optical imaging and MRI, however, the reporter
groups are bulky molecules or complexes, which have a dominant
influence on the pharmacokinetic properties of the ligand-reporter
construct. In fact, the predominant number of NIRF and MRI studies
to date has targeted endovascular or extracellular rather than intra-
cellular targets. Several concepts have been developed to enhance
cell penetration of such reporter groups (Lewin et al. 2000; Hoehn
et al. 2002; Poduslo et al. 2002); however, they have not been used
for in vivo labeling in the intact animal but rather to label harvested
cells, which were reinjected and monitored in vivo. For molecular
imaging of intracellular targets, future developments are urgently
needed.
3.8 Conclusion
Today, in vivo imaging has become indispensable in biomedical re-

search. Imaging approaches are applied at essentially all steps of
drug discovery and development, from the validation of a potential
disease-relevant drug target to the clinical drug evaluation. The re-
cent development of target-specific or molecular imaging techniques
has certainly added a new dimension: structural and functional infor-
mation can now be complemented by data on molecular biology col-
lected in the intact organism. Molecular imaging thus constitutes a
bridge between the holistic aspects covered by structural and func-
tional imaging, providing information on the disease phenotype, and
the reductionistic, mechanistic concepts relating pharmacological ac-
tivities to molecular interactions (drug-target interaction, signal
transduction pathways).
The imaging field is still rapidly evolving and many applications
that seem beyond reach today will be feasible within the next de-
cade. The success of molecular imaging applications will critically
depend on probe development. The objectives are to develop highly
specific and sensitive reporters, with favorable PK properties in or-
der to efficiently reach the target without affecting its function. De-
velopment in imaging technologies will improve the robustness of
the techniques, the ease of operation, and the tools for quantitative
data analysis (concentrations, activities, fluxes instead of signal in-
tensities). In a not-too-distant future, many of these technologies
will leave the specialist lab and will be applied in the in vivo phar-
macology lab; this will, ultimately, lead to an inflation of biomedi-
cal imaging applications.
References
Agdeppa ED, Kepe Y, Liu J, Flores-Torres S, Satyamurthy N, Petrie A, Cole

GM, Small GW, Huang SC, Barrio JR (2001) Binding characteristics of
radiofluorinated 6-dialkylamino-2-naphthylethylidene derivatives as posi-
tron emission tomography imaging probes for beta-amyloid plaques in
Alzheimer's disease. J Neurosci 21 :RC189
Aguayo JB, Blackband SJ, Schoeninger J, Mattingly MA, Hinterrnann M
(1986) MRI of single cells. Nature 322: 190-191
68 M. Rudin et al.
Beauregard M, Leroux JM, Bergman S, Arzoumanian Y, Beaudoin G, Bour-

gouin P, Stip E (1998) The functional neuroanatomy of major depression:
an fMRI study using an emotional activation paradigm. NeuroReport
9:3253-3258
Becker A, Hessenius C, Licha K, Ebert B, Sukowski U, Semmler W, Wie-
denmann B, Grotzinger C (2001) Receptor-targeted optical imaging of tu-
mors with near-infrared fluorescent ligands. Nat Biotechnol 19:327-331
Beckmann N, Hof RP, Rudin M (2000) The role of magnetic resonance
imaging and spectroscopy in transplantation: from animal models to man.
NMR Biomed 13:329-348
Beckmann N, Mueggler T, Allegrini PR, Laurent D, Rudin M (2001) From
anatomy to the target: contributions of magnetic resonance imaging to
preclinical pharmaceutical research. Anat Rec 265:85-100
Beckmann N, Cannet C, Fringeli-Tanner M, Baumann D, Pally C, Bruns C,
Zerwes HG, Andriambeloson E, Bigaud M (2003) Analysis of a rat mod-
el of kidney transplantation by morphological and functional MRI:
Search for early markers of allograft chronic rejection. Magn Reson Med
49:459-467
Benaron DA, Con tag PR, Contag CH (1997) Imaging brain structure and
function, infection and gene expression in the body using light. Phil os
Trans R Soc Lond B BioI Sci 352:755-761
Benveniste H, Hedlund LW, Johnson GA (1992) Mechanism of detection of
acute cerebral ischemia in rats by diffusion-weighted magnetic resonance
microscopy. Stroke 23:746-754
Benveniste H, Einstein G, Kim KR, Hulette C, Johnson GA (1999) Detec-
tion of neuritic plaques in Alzheimer's disease by magnetic resonance
microscopy. Proc Natl Acad Sci USA 96:14079-14084
Birchler M, Neri G, Tarli L, Halin C, Viti F, Neri D (1999) Infrared photo-
detection for the in vivo localisation of phage-derived antibodies directed
against angiogenic markers. J Immunol Methods 231 :239-244
Blankenberg FG, Katsikis PD, Tait JF, Davis RE, Naumovski L, Ohtsuki K,
Kopiwoda S, Abrams MJ, Darkes M, Robbins RC, Maecker HT, Strauss
HW (1998) In vivo detection and imaging of phosphatidylserine expres-
sion during programmed cell death. Proc Nat! Acad Sci USA 95:6349-6354
Bolo NR, Hode Y, Nedelec JF, Laine E, Wagner G, Macher JP (2000) Brain
pharmacokinetics and tissue distribution in vivo of fluvoxamine and
fluoxetine by fluorine magnetic resonance spectroscopy. Neuropsycho-
pharmacol 23:428-438
Botteron KN, Raichle ME, Drevets WC, Heath AC, Todd RD (2002) Volu-
metric reduction in left subgenual prefrontal cortex in early onset depres-
sion. BioI Psychiatry 51 :342-344
Brasch R, Pham C, Shames D, Roberts T, van Dijke K, van Bruggen N,
Mann J, Ostrowitzki Melnyk 0 (1997) Assessment of tumor angiogenesis
using macromolecular MRE imaging contrast media. J Magn Reson
Imaging 7:68-74
Bremer C, Tung CH, Weissleder R (2001) In vivo molecular target assess-

ment of matrix metalloproteinase inhibition. Nat Med 7:743-748
Bulte JW, Zhang S, van Gelderen P, Herynek V, Jordan EK, Duncan ID,
Frank JA (1999) Neurotransplantation of magnetically labeled oligoden-
drocyte progenitors: magnetic resonance tracking of cell migration and
myelination. Proc Natl Acad Sci USA 96:15256-15261
Carlsen H, Moskau JO, Fromm SH, Blomhoff R (2002) In vivo imaging of
NF-KB activity. J Immunol 168:1441-1446
Chen J, Tung CH, Mahmood U, Ntziachristos V, Gyurko R, Fishman MC,
Huang PL,Weissleder R (2002) In vivo imaging of proteolytic activity in
atherosclerosis. Circulation 105:2766-2771
Chen YCI, Galpem WR, Brownell AL, Matthews RT, Bogdanov M, Isacson
0, Keltner JR, Beal MF, Rosen BR, Jenkins BG (1997) Detection of
dopaminergic neurotransmitter actIvIty using pharmacologic
MRI:correlation with PET, microdialysis and behavioural data. Magn Re-
son Med 38:389-398
Chevenert TL, McKeever PE, Ross BD (1997) Monitoring early response of
experimental brain tumors to therapy using diffusion magnetic resonance
imaging. Clin Cancer Res 3:1457-1466
Chevenert TL, Stegman LD, Taylor JMG, Robertson PL, Greenberg HS, Re-
hemtulla A, Ross BD (2000) Diffusion magnetic resonance imaging: an
early surrogate marker of therapeutic efficacy in brain tumors. J Natl
Cancer Inst 92:2029-2036
Ciana P, Raviscioni M, Mussi P, Vegeto E, Que I, Parker MM, Lowik C,
Maggi A (2003) In-vivo imaging of transcriptionally active estrogen re-
ceptors. Nat Med 9:82-86
Clement 0, Pradel C, Siauve N, Frouin F, Bruneteau G, Kahn E, Frija G,
Cuenod CA (2002) Assessing perfusion and capillary permeability
changes induced by a VEGF inhibitor in human tumor xenografts using
macromolecular MR imaging contrast media. Acad Radiol 9(Suppl
2):S328-S329
Cohen SM, Werrmann JG, Rasmusson GH, Tanaka WK, Malatesta PF, Pra-
hal ada S, Jacobs JG, Harris G, Nett TM (1995) Comparison of the ef-
fects of anew specific azasteroid inhibitor of 5a-reductase on canine hy-
perplastic prostate. Suppression of prostatic DHT correlated with prostate
regression. Prostate 26:55-72
Contag CH, Spilman SD, Contag PR, Oshiro M, Eames B, Dennery P, Steven-
son DK, Benaron DA (1997) Visualizing gene expression in living mam-
mals using a bioluminescent reporter. Photochem Photobiol 66:523-531
Costa GL, Sandora MR, Nakajima A, Nguyen EV, Taylor-Edwards C, Slavin
AJ, Contag CH, Fathman CG, Benson JM (2001) Adoptive immunother-
apy of experimental autoimmune encephalomyelitis via T cell delivery of
the IL-12 p40 subunit. J Immunol 167:2379-2387
Dousset V, Delalande C, Ballarino L, Quesson B, Seilhan D, Coussemacq
M, Thiaudiere E, Brochet B, Canioni P, Caille JM (1999) In-vivo macro-
70 M. Rudin et al.
phage activity imaging in central nervous system detected by magnetic

resonance. Magn Reson Med 41:329-333
Drevs J, Muller-Driver R, Wittig C, Fuxius S, Esser N, Hugenschmidt H,
Konerding MA, Allegrini PR, Wood J, Hennig J, Unger C, Marme D
(2002) PTK787/ZK 222584, a specific vascular endothelial growth fac-
tor-receptor tyrosine kinase inhibitor, affects the anatomy of the tumor
vascular bed and the functional vascular properties as detected by dy-
namic enhanced magnetic resonance imaging. Cancer Res 62:4015-4022
Duvvuri U, Poptani H, Felman M, Nadal-Desbarats L, Gee MS, Lee WMF,
Reddy R, Leigh JS, Glickson JD (2001) Quantitative Tlr magnetic reso-
nance imaging of RIF-l tumors in-vivo: DETECTION of early response
to cyclophosphamide therapy. Cancer Res 61:7747-7753
Eliceiri BP, Cheresh DA (2001) Adhesion events in angiogenesis. CUff Opin
Cell Bioi 13(5):563-568
Evelhoch JL, Gillies RJ, Karczmar GS, Koutcher JA, Maxwell RJ, Nalcioglu
0, Raghunand N, Ronen SM, Ross BD, Swartz HM (2000) Applications
of magnetic resonance in model systems: cancer therapeutics. Neoplasia
2:152-165
Fischman AJ, Alpert NM, Rubin RH (2002) Pharmacokinetic imaging: a
noninvasive method. Clin Pharmacokinet 41:581-602
Furman-Haran E, Margalit R, Grobgeld D, Degani H (1996) Dynamic con-
trast-enhanced magnetic resonance imaging reveals stress-induced angio-
genesis in MCF7 human breast tumors. Proc Nat! Acad Sci USA
93:6247-6251
Fuxe K, Ungerstedt U (1976) Antiparkinsonian drugs and dopaminergic
neostriatal mechanisms: studies in rats with unilateral 6-hydrox ydopa-
mine (6-0H-DA)-induced degeneration of the nigro-neostriatal DA path-
way and quantitative recording of rotational behaviour. Pharmacol Ther
[B] 2:41-47
Green LA, Yap CS, Nguyen K, Barrio JR, Namavari M, Satayamurthy N,
Phelps ME, Sandgren EP, Hershman HR, Gambhir SS (2002) Indirect
monitoring of endogenous gene expression by positron emission tomogra-
phy (PET) imaging of reporter gene expression in transgenic mice. Mol
Imaging Bioi 4:71-81
Griffiths JR, Glickson JD (2000) Monitoring pharmacokinetics of anticancer
drugs: non-invasive investigation using magnetic resonance spectroscopy.
Adv Drug Deliv Rev 41:75-89
Haeusler A, Allegrini PR, Biollaz M, Batzl C, Scheidegger E, Bhatnagar AS
(1996) CGP 53153: a new potent inhibitor of Sa-reductase. J Steroid Bio-
chern Mol Bioi 58:187-195
Hoehn M, Kustermann E, Blunk J, Wiedermann D, Trapp T, Wecker S,
Focking M, Arnold H, Hescheler J, Fleischmann BK, Schwindt W,
Buhrle C (2002) Monitoring of implanted stem cell migration in vivo: a
highly resolved in vivo magnetic resonance imaging investigation of ex-
perimental stroke in rat. Proc Nat! Acad Sci USA 99:16267-16272
Hood JD, Bednarski M, Frausto R, Guccione S, Reisfeld RA, Xiang R,

Cheresh DA (2002) Tumor regression by targeted gene delivery to the
neovasculature. Science 296:2404-2407
Iyer M, Wu L, Carey M, Wang Y, Smallwood A, Gambhir SS (2001) Two-
step transcriptional amplification as a method for imaging reporter gene
expression using weak promoters. Proc Natl Acad Sci USA 98: 14595-
14600
Kaim A, Wi scher T, O'Reilly T, Jundt G, Froehlich J, Von Schulthess GK,
Allegrini PR (2002) MR imaging with ultrasmall superparamagnetic iron
oxide particles in experimental soft-tissue infections in rats. Radiology
225: 808-814
Knopp MV, Weiss E, Sinn HP, Mattern J, Junkermann H, Radeleff J, Mage-
ner A, Brix G, Delorme S, Zuna I, van Kaick G (1999) Pathophysiologi-
cal basis of contrast enhancement in breast tumors. J Magn Reson Imag-
ing 10:260-266
Lalonde R, Dumont M, Staufenbiel M, Sturchler-Pierrat C, Strazielle C
(2002) Spatial learning, exploration, anxiety, and motor coordination in
female APP23 transgenic mice with the Swedish mutation. Brain Res
956:36-44
Langstrom B, Kihlberg T, Bergstrom M, Antoni G, Bjorkman M, Forngren
BH, Forngren T, Hartvig P, Markides K, Yngve U, Ogren M (1999)
Compounds labelled with short-lived beta(+)-emitting radionuclides and
some applications in life sciences. The importance of time as a parame-
ter. Acta Chern Scand 53:651-669
Laxman B, Hall DE, Bhojani MS, Hamstra DA, Chevenert TL, Ross BD,
Rehemtulla A (2002) Noninvasive real-time imaging of apoptosis. Proc
Nat! Acad Sci USA 99:16551-16555
Lewin M, Carlesso N, Tung CH, Tang XW, Cory D, Scadden DT, Weissle-
der R (2000) Tat peptide-derivatized magnetic nanoparticles allow in vivo
tracking and recovery of progenitor cells. Nat Biotechnol 18:410-414
Li CPK, Bednarski MD (2002) Vascular-targeted molecular imaging using
functionalized polymerized vesicles. J Magn Reson Imaging 16:388-393
Li CPK, Guccione S, Bednarski MD (2002) Combined vascular targeted
imaging and therapy: A paradigm for personalized treatment. J Cell Bio-
chern (Suppl) 39:605
Louie AY, Huber MM, Ahrens ET, Rothbacher U, Moats R, Jacobs RE, Fra-
ser SE, Meade TJ (2000) In vivo visualization of gene expression using
magnetic resonance imaging. Nat Biotechnol 18:321-325
Luker GD, Sharma V, Pica CM, Dahlheimer JL, Li W, Ochesky J, Ryan CE,
Piwnica-Worms H, Piwnica-Worms D (2002) Noninvasive imaging of
protein-protein interactions in living animals. Proc Nat! Acad Sci USA
99:6961-6966
Mahmood U, Tung CH, Bogdanov A, Weissleder R (1999) Near-infrared op-
tical imaging of protease activity for tumor detection. Radiology
213: 866-870
72 M. Rudin et al.
Mandl S, Schimmelpfennig C, Edinger M, Negrin RS, Contag CH (2002)

Understanding immune cell trafficking patterns via in vivo biolumines-
cence imaging. J Cell Biochem (Suppl) 39:239-248
Massoud TF, Gambhir SS (2002) Molecular imaging in living subjects: seeing
fundamental biological processes in a new light. Genes Dev 17:545-580
McLaren DC, Gambhir SS, Satayamurthy N, Barrio JR, Scharfstein S, Toyo-
kuni T, Wu L, Berk A, Cherry SR, Phelps ME, et al (1999) Repetitive,
non-invasive imaging of the dopamine D2 receptor as a reporter gene in
living animals. Gene Ther 6:785-791
Mueggler T, Sturchler-Pierrat C, Baumann D, Rausch M, Staufenbiel M, Ru-
din M (2002a) Compromised hemodynamic response in amyloid precur-
sor protein transgenic mice. J Neurosci 15:7218-7224
Mueggler T, Sturchler-Pierrat C, Baumann D, Rausch M, Rudin M (2002b)
Compromised hemodynamic response in APP transgenic mice detected
by functional MRI. Proceedings of the 10th ISMRM, Honolulu, p 402
Nakajima A, Seroogy CM, Sandora MR, Tamer IH, Costa GL, Taylor-Ed-
wards C, Bachmann MH, Contag CH, Fathman CG (2001) Antigen-spe-
cific T-cell-mediated gene therapy in collagen-induced arthritis. J Clin In-
vest 107: 1293-1301
Namavari M, Barrio JR, Toyokuni T, Gambhir SS, Cherry SR, Herschman
HR, Phelps ME, Satayamurthy N (2000) Synthesis of 8-[18F] fluorogua-
nine derivatives: in-vivo probes for imaging gene expression with PET.
Nucl Med Bioi 27:157-162
Nicholson JK, Connelly J, Lindon JC, Holmes E (2002) Metabonomics: a
platform for studying drug toxicity and gene function. Nat Rev Drug Dis-
cov 1:153-161
Ntziachristos V, Weissleder R (2002) Charge-coupled-device based scanner
for tomography of fluorescent near-infrared probes in turbid media. Med
Phys 29:803-809
Ntziachristos V, Tung CH, Bremer C, Weissleder R (2002) Fluorescence mo-
lecular tomography resolves protease activity in vivo. Nat Med 8:757-760
Dh C, Murray B, Bhattacharya N, Holland D, Tatton WG (1994) (-)-Depre-
nyl alters the survival of adult murine facial motor neurons after axot-
omy: increases in vulnerable C57BL strain but decreases in motor neuron
degeneration mutants. J Neurosci Res 38:64-74
Petrovsky A, Schellenberger E, Josephson L, Weissleder R, Bogdanov A
(2003) Near-Infrared fluorescent imaging of tumor apoptosis. Cancer Res
63: 1936-1942
Poduslo JF, Wengenack TM, Curran GL, Wisniewski T, Sigurdsson EM, Ma-
cura SI, Borowski BJ, Jack CR (2002) Molecular targeting of Alzhei-
mer's amyloid plaques for contrast-enhanced magnetic resonance imag-
ing. Neurobiol Dis 11:315-329
Price DL, Tanzi RE, Borchelt DR, Sisodia SS (1998) Alzheimer's disease:
genetic studies and transgenic models. Annu Rev Genet 32:461-493
Rausch M, Sauter A, Frohlich J, Neubacher U, Radii EW, Rudin M (2001)

Dynamic pattern of USPIO enhancement can be observed in macro-
phages after ischemic brain damage. Magn Reson Med 46: 10 18-1021
Rausch M, Baumann D, Neubacher U, Rudin M (2002) In-vivo visualization
of phagocytotic cells in rat brains after transient ischemia by USPIO.
NMR Biomed 15:278-283
Rausch M, Hiestand P, Baumann D, Cannet C, Rudin M (2003) MRI based
monitoring of inflammation and tissue damage in acute and chronic re-
lapsing EAE. Magn Reson Med 50:309-314
Ray P, Pimenta H, Paulmurugan R, Berger F, Phelps ME, Iyer M, Gambhir
SS (2002) Noninvasive quantitative imaging of protein-protein interac-
tions in living subjects. Proc Natl Acad Sci USA 99:3105-3110
Reese T, Bjelke B, Porszasz R, Baumann D, Bochelen D, Sauter A, Rudin
M (2000) Regional brain activation by bicuculline visualized by func-
tional magnetic resonance imaging. Time-resolved assessment of bicucul-
line-induced changes in local cerebral blood volume using an intravascu-
lar contrast agent. NMR Biomed 13:43-49
Rudin M, Weissleder R (2003) Molecular imaging in drug discovery and de-
velopment. Nat Rev Drug Discov 2:123-131
Rudin M, Briner U, Doepfner W (1988) Quantitative magnetic resonance
imaging of estradiol-induced pituitary hyperplasia in rats. Magn Reson
Med 7:285-291
Rudin M, Beckmann N, Porszasz R, Reese T, Bochelen D, Sauter A (1999)
In vivo magnetic resonance methods in pharmaceutical research: current
status and perspectives. NMR Biomed 12:69-97
Salazar DE, Fischman AJ (1999) Central nervous system pharmacokinetics
of psychiatric drugs. J Clin Pharmacol (Suppl): lOS-12S
Sauter A, Rudin M (1987) Effects of calcium anta¥onists on high-energy
phosphates in ischemic rat brain measured by 3 P NMR spectroscopy.
Magn Reson Med 4:1-8
Sauter A, Rudin M, Wiederhold KH, Hof RP (1989) Cerebrovascular, bio-
chemical, and cytoprotective effects of isradipine in laboratory animals.
Am J Med 86 (Suppl 4A):134-146
Sauter A, Reese T, Porszasz R, Baumann D, Rausch M, Rudin M (2002)
Recovery of function in cytoprotected cerebral cortex in rat stroke model
assessed by functional MRI. Magn Reson Med 47:759-765
Scheffold C, Kornacker M, Scheffold YC, Contag CH, Negrin RS (2002)
Visualization of effective tumor targeting by CD8+ natural killer T cells
redirected with bispecific antibody F(ab')(2)HER2xCD3. Cancer Res
62:5785-5791
Schellenberger EA, Hogemann D, Josephson L, Weissleder R (2002) Annex-
in V-CLIO: a nanoparticle for detecting apoptosis by MRI. Acad Radiol
9(Suppl 2):S31O-S311
Shoghi-Jadid K, Small GW, Agdeppa ED, Kepe V, Ercoli LM, Siddarth P,
Read S, Satyamurthy N, Petric A, Huang SC, Barrio JR (2002) Localiza-
74 M. Rudin et al.
tion of neurofibrillary tangles and beta-amyloid plaques in the brains of

living patients with Alzheimer disease. Am J Geriatr Psychiatry 10:24-35
Siegel RA, Tolcsvai L, Rudin M (1988) Partial inhibition of the growth of
transplanted Dunning rat prostate tumors with the long-acting somatosta-
tin analog Sandostatin (SMS 201-995). Cancer Res 48:4651-4655
Sipkins DA, Cheresh DA, Kazemi MR, Nevin LM, Bednarski MD, Li KC
(1998) Detection of tumor angiogenesis in vivo by alphav-beta3-targeted
magnetic resonance imaging. Nat Med 4:623-626
Sipkins DA, Gijbels K, Tropper FD, Bednarski M, Li KC, Steinman L
(2000) ICAM-I expression in autoimmune encephalitis visualized using
magnetic resonance imaging. J Neuroimmunol 104:1-9
Smith AD, Jobst KA (1996) Use of structural imaging to study the progres-
sion of Alzheimer's disease. Br Med Bull 52:575-586
Strauss WL, Layton ME, Hayes CE, Dager SR (1997) 19F magnetic resonance
spectroscopy investigation in vivo of acute and steady-state brain fluvoxa-
mine levels in obsessive-compulsive disorder. Am J Psychiatry 154:516-522
Sturchler-Pierrat C, Abramowski D, Duke M, Wiederhold KH, Mist! C,
Rothacher S, Lederrnann B, Burki K, Frey P, Paganetti PA, Waridel C,
Calhoun ME, Jucker M, Probst A, Staufenbiel M, Sommer B (1997) Two
amyloid precursor protein transgenic mouse models with Alzheimer dis-
ease-like pathology. Proc Natl Acad Sci USA 94:13287-13292
Tropres I, Grimault S, Vaeth A, Grillon E, Julien C, Payen JF, Lamalle L,
Decorps M (2001) Vessel size imaging. Magn Reson Med 45:397-408
Tung CH, Mahmood U, Bredow S, Weissleder R (2000) In vivo imaging of
proteolytic enzyme activity using a novel molecular reporter. Cancer Res
60:4953-4958
Tung CH, Lin Y, Moon WK, Weissleder R (2002) A receptor-targeted near-
infrared fluorescence probe for in vivo tumor imaging. Chembiochem
3:784-786
Turetschek K, Preda A, Floyd E, Shames DM, Novikov V, Roberts TP,
Wood JM, Fu Y, Carter WO, Brasch RC (2002) MRI monitoring of tu-
mor response to a novel VEGF tyrosine kinase inhibitor in an experimen-
tal breast cancer model. Acad Radiol 9(Suppl 2):S519-S520
Van den Abbeele AD, Badawi RD (2002) Use of positron emission tomogra-
phy in oncology and its potential role to assess response to imatinib me-
sylate therapy in gastrointestinal stromal tumors (GISTs). Europ J Cancer
38(Suppl 5):S60-65
van Dijke C, Brasch RC, Roberts TP, Weidner N, Mathur A, Shames DM,
Desmar F, Lang P, Schwickert HC (1996) Mammary carcinoma model:
correlation of macromolecular contrast-enhanced MR imaging character-
ization of tumor microvasculature and histologic capillary density. Radi-
ology 198:813-818
van Royen EA, Verhoeff NP, Meylaerts SA, Miedema AR (1996) Indium-
lll-DTPA-octreotide uptake measured in normal and abnormal pituitary
glands. J Nue! Med 37:1449-1451
Wadghiri YZ, Sigurdsson E, Tang C, Turnbull D (2002) MR micro-Imaging

of contrast tagged amyloid plaques in transgenic mouse models of Alz-
heimer's disease. Proceedings of 10th ISMRM Meeting, p 388
Weissleder R, Ntziachristos V (2003) Shedding light onto live molecular tar-
gets. Nat Med 9:123-128
Weissleder R, Tung CH, Mahmood U, Bogdanov A (J 999) In vivo imaging
of tumors with protease-activated near-infrared fluorescent probes. Nat
Biotechnol 17:375-378
Weissleder, R, Moore A, Mahmood U, Bhorade R, Benveniste H, Chiocca
EA, Basilion JP (2000) In vivo magnetic resonance imaging of transgene
expression. Nat Med 6:351-355
Wetterwald A, van der Pluijm G, Que I, Sijmons B, Buijs J, Karperien M,
Lowik CW, Gautschi E, Thalmann GN, Cecchini MG (2002) Optical
imaging of cancer metastasis to bone marrow: a mouse model of mini-
mal residual disease. Am J Pathol 160: 1143-1153
Wolf W, Presant CA, Waluch V (2000) 19F-MRS studies of fluorinated
drugs in humans. Adv Drug Deliv Rev 41:55-74
Xu H, Li SJ, Bodurka J, Zhao X, Zheng XX, Stein AA (2000) Heroin-in-
duced neuronal activation in rat brain assessed by functional MRI. Neu-
roReport 11: 1085-1092
Zhao M, Beauregard DA, Loizou L, Davletov B, Brindle KM (2001) Non-
invasive detection of apoptosis using magnetic resonance imaging and a
targeted contrast agent. Nat Med 7:1241-1244
4 Internal Dose Assessment -
Extrapolation from Animal Species
to the Human
M.G. Stabin
4.1 Introduction................................. 77
4.2 Basic Equations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.3 Main Equation ............................... 79
4.4 Analysis of Kinetic Data - General Concepts . . . . . . . . . . . . 82
4.5 Analysis of Kinetic Data - Collecting Animal Data . . . . . . . . 83
4.6 Extrapolation of Results ......................... 85
4.7 Kinetic Analysis .............................. 87
4.8 Application of Dosimetric Models . . . . . . . . . . . . . . . . . . . 88
4.9 Development of MIRDOSE Codes. . . . . . . . . . . . . . . . . . . 90
4.10 Conclusions ................................. 91
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.1 Introduction
Many different radionuclides are in use or being considered for use

in radiation therapy with radionuclides. Examples include the tradi-
tional uses of radioiodine for treatment of thyroid disorders (hy-
perthyroidism, thyroid cancer), the use of monoclonal antibodies,
peptides, and other agents against specific types of cancer, and the
use of various bone seeking agents used in the palliation of pain for
bony metastases. Estimation of radiation doses for any of these ap-
plications involves several complex phases of analysis. Many stan-
dard models and techniques are in use, but it is also a time of rapid
78 M.G. Stabin
development of new techniques and approaches. It is indisputable

that even the best current methods for calculating radiation dose
give only estimates of the radiation dose to patients' organs, marrow,
tumors, blood vessel walls, etc. Use of animal data to predict human
dosimetry introduces additional uncertainties into the analyses, as no
truly reliable method exists for making this extrapolation. Human
dose estimates based on animal data are considered to be reasonable
approximations to be used for proceeding with dose estimates based
on human data, which are ultimately used to evaluate in the safety
and efficacy evaluations of radiopharmaceuticals. Nonetheless, the
calculation of radiation dose based on animal data continues to be
an important element of the radiopharmaceutical approval process.
4.2 Basic Equations
To define the task of calculating internal doses, we must define the

quantities we wish to estimate. The principal quantity of interest in
internal dosimetry is the absorbed dose, or the dose equivalent. Ab-
sorbed dose (D) is defined by the International Commission on Ra-
diation Units and Measurements (ICRU 1980) as:
de
D=-
dm
where de is the mean energy imparted by ionizing radiation to matter

of mass dm. The units of absorbed dose are typically erg/g or Jlkg. The
special units are rad (l00 erg/g) or the gray (Gy) (l Jlkg= 100 rad=
104 erg/g). The dose equivalent (H) is the absorbed dose multiplied
by a 'quality factor' (Q) or more recently, the 'radiation weighting
factor' (WR), which accounts for the effectiveness of different types
of radiation in causing biological effects:
H=DQ
Because the quality factor is in principle dimensionless, the pure units

of this quantity are the same as absorbed dose (i.e., erg/g or Jlkg).
However, the special units have unique names, specifically, the rem
and sievert (Sv). Values for the quality factor have changed as new
Internal Dose Assessment 79
Table 1. Quality factors recommended in ICRP 30
Alpha particles 20
Beta particles (+/-) I
Gamma rays I
X-rays I
information about radiation effectiveness has become available. Cur-

rent values, recommended by the International Commission on Ra-
diological Protection (ICRP 1979), are given in Table 1.
These values, particularly for alpha particles and low-energy elec-
trons, are still quite uncertain and are subject to change as new
radiobiological information becomes available. The quantity dose
equivalent was originally derived for use in radiation protection pro-
grams. The development of the effective dose equivalent (EDE) by
the ICRP in 1979, and the effective dose (ED), in 1991, however, al-
lowed nonuniform internal doses to be expressed as a single value,
representing an equivalent whole-body dose.
4.3 Main Equation

In order to estimate absorbed dose for all significant tissues, one
must determine for each tissue the quantity of energy absorbed per
unit mass. This yields the quantity 'absorbed dose,' if expressed in
proper units, and can be extended to a calculation of dose equivalent
if desired. What quantities are then needed to calculate the two key
parameters, energy and mass? We can show a generic equation for
the absorbed dose rate in our object as:
iJ = ----'----
m
iJ = absorbed dose rate (rad/hr or Gy/sec)

A = activity (,uCi or MBq)
n = number of radiations with energy E emitted per nuclear transition
E = energy per radiation (Me V)
rjJ = fraction of energy absorbed in the target
80 M.G. Stabin
m = mass of target region (g or kg)

k = proportionality constant (rad-g/,uCi-hr-MeVor Gy-kg/MBq-sec-
MeV)
It is extremely important that the proportionality constant be properly

calculated and applied. The results of our calculation will be useless
unless the units within are consistent and they correctly express the
quantity desired. The application of quality factors to this equation
to calculate the dose equivalent rate is a trivial matter; for most of this
chapter, we will consider only absorbed doses for discussion purposes.
The investigator is not usually interested only in the absorbed dose
rate; more likely an estimate of total absorbed dose from an adminis-
tration is desired. In this equation, the quantity activity (nuclear transi-
tions per unit time) causes the outcome of the equation to have a time
dependence. In order to calculate cumulative dose, the time integral of
the activity must be calculated. Regardless of the shape of the time
activity curve, its integral, however obtained, will have units of transi-
tions (activity, which is transitions per unit time, multiplied by time).
Therefore, the equation for cumulative dose would be:
D = ---'-----
m
D =absorbed dose (rad or Gy)

A =cumulated activity (,uCi-h or MBq-s)
The quantity cumulated activity CA) gives the area under a time-ac-
tivity curve:
Activity
A
Time
If activity is in units of Bq and time is in units of seconds, A will

have units of Bq-s. This is a measure of the number of disintegra-
tions that have occurred in a source region over time - Bq is a num-
ber of disintegrations per second, thus A has units of disintegrations.
If activity is in units of ,uCi and time is in hours, the principle is the
same; 1 ,uCi-h is equivalent to 1.33x108 disintegrations.
The basic internal dose equation can be thought of as having two
components: a biological component and a physical component.
This is reflected in the two parts of the basic Medical Internal Ra-
diation Dose (MIRD) dose equation (Loevinger et al. 1988):
D=AS
D = Ao'S
D
-=,S
Ao
where
D =the absorbed dose (rad or Gy)
A =the cumulated activity (,uCi-h or Bq-s)
S = the S-value (rad!,uCi-hr or Gy/Bq-s)
Ao = the administered activity (,uCi or Bq)
, =the residence time (h or s)
The biological component (A or ,) includes the parameters describ-

ing the distribution of the pharmaceutical in the body and its reten-
tion kinetics. The physical component (S) includes the parameters
describing the types and energies of the radionuclide decay, the ab-
sorption of energy in the body and the masses of the body organs.
In general, the physics of a system, especially a standardized system
such as for reference adult man or woman, are determined and ac-
cessed through stored computer files. The biological parameters,
however, must be determined for each new compound studied. This
is usually accomplished through studies involving animals or hu-
mans, in which the distribution, retention, and excretion are quanti-
fied and fit to a model. In general, the uncertainties in the biological
parameters are much greater than those in the physical parameters,
even when one considers that the dose reported is an average for a
82 M.G. Stabin
population. The measured values in a kinetic study, and their appli-

cability to different individuals, are probably the source of the great-
est uncertainty in the model.
4.4 Analysis of Kinetic Data - General Concepts
To design and execute a good kinetic study, one needs to collect the
correct data, enough data, and express the data in the proper units.
The basic data needed are the fraction (or percentage) of adminis-
tered activity in important source organs and excreta samples. It is
very important, in either animal or human studies, to take enough
samples to characterize both the distribution and retention of the
radiopharmaceutical over the course of the study. The following cri-
teria are essential:
- Catching the early peak uptake and rapid washout phase

- Covering at least three effective half times of the radiopharmaceu-
tical
- Collecting at least two time points per clearance phase
- Accounting for 100% of the activity at all times
- Accounting for all major paths of excretion (urine, feces, exhala-
tion, etc.)
Some knowledge of the expected kinetics of the pharmaceutical are

needed for a good study design. For example, the spacing of the
measurements and the time of the initial measurement will be
greatly different if we are studying a Tc-99m-Iabeled renal agent
which is 95% cleared from the body in 180 min or an I-131-labeled
antibody which clears about 80% in the first day and the remaining
20% over the next 2 weeks. A key point which researchers often
overlook is the characterization of excretion. Very often, the excre-
tory organs (intestines, urinary bladder) are the organs that receive
the highest absorbed doses, as 100% of the activity (minus decay)
will eventually pass through one or both of these pathways at differ-
ent rates. If excretion is not quantified, the modeler must assume
that the compound is retained in the body and removed only by
radioactive decay. For very short-lived radionuclides, this may not
be a problem and in fact may be quite accurate. For moderately

long-lived nuclides, this can cause an overestimate of the dose to
most body organs and an underestimate of the dose to the excretory
organs, perhaps significantly. For very long-lived nuclides (e.g., 14C,
4oK), this assumption is usually not realistic and will result in the
calculation of very large committed doses.
Some steps can be taken to optimize the timing of the data points
to minimize the potential error in the final estimates of the organ
residence times. The MIRD Committee has published a document
that addresses these considerations (Siegel et al. 1999). The docu-
ment outlines the proper techniques for data quantification and for
appropriate temporal sampling. The authors show how to use the
conjugate view technique to acquire quantitative data for dosimetry
analyses, including proper choice of source and background regions,
with corrections for overlapping source regions, background, and
scatter. The use of single photon emission computed tomography
(SPECT) and positron emission tomography (PET) techniques also
are discussed. Quantitative methods for analyzing blood and excreta
samples are described. On the subject of temporal sampling, they
demonstrate that two or three time points per phase (either uptake or
clearance) are needed to adequately describe the kinetics. They also
show graphically the amount of error in A (or r) that occurs from
neglecting the wash-in phase in an organ or not adequately assessing
the wash-out phase. The authors provide several valuable examples
for many categories of calculation, which makes the document use-
ful to the practitioner. This publication provides a useful aid in de-
signing kinetic studies; however, in each individual case it is the re-
sponsibility of the investigator to adequately describe the time-activ-
ity curves in all source organs which have a significant uptake of the
radiopharmaceutical, the organs involved in the excretion of the
compound, and tissues in the remainder of the body.
4.5 Analysis of Kinetic Data - Collecting Animal Data
Several concerns are involved in the successful execution of the precli-

nical phase of research with a new radiopharmaceutical. The first is the
choice of an animal species. This is driven by a number of concerns,
84 M.G. Stabin
including size, cost, suitability to the study, ease or difficulty in han-

dling, and other issues. For example, rats do not have a gallbladder,
so they will not be useful in characterizing the dosimetry of hepatobili-
ary agents. Certain species may not express the type of cancer that you
are interested in studying. The placenta is very different in different
animals; these differences should be appreciated before designing a
study to predict placental crossover of radionuclides based on animal
results. Sometimes, it may not be possible to know ahead of time
which animals will or will not characterize the distribution well; exam-
ples of surprises are discussed in the section below on experience with
preclinical data. The bottom line is that the researcher should exercise
as much discretion as possible before the study to avoid surprises.
Then, one must choose an analytical method. In autoradiography
studies, animals are sacrificed and frozen into a block of carboxy-
methyl cellulose (CMC), slices are cut with a cryostat microtome,
where the knife an cuts approximately 30 J.lm slice of the animal,
and the sections are then freeze dried and attached to X-ray film for
exposure. Exposure of the film demonstrates areas of uptake and con-
centration. Suitable calibration can make such analyses quantitative.
Necropsy studies involve sacrifice of animals at different time
points, freezing of the tissues, and dissection of individual organs
and tissues from the carcass. These tissues are then assayed for radio-
activity content, typically using scintillation detectors. Recent ad-
vances in small animal imaging technologies have suggested that a
change from the historical limitations of autoradiography or necropsy
studies may be possible. The use of microPET and microSPECT meth-
ods, with region of interest (ROI) analysis of images, as is done in clin-
ical studies in humans, is envisioned. This is in the experimental stage
at present, and may not become routine even in well-equipped medical
centers for some time to come. Augmenting the evaluation of whole-
body and organ uptake is an evaluation of the pharmaceutical excre-
tion, usually done with metabolism cages, which are simply small sin-
gle animal cages in which the animal is housed after the injection of
the radioactivity. The excreta (urine and feces) are collected periodi-
cally from the cage. Some type of paper on the floor of the cage,
for example filter paper, will help to limit the cross contamination
of urine and fecal activity. A cage wash at each collection time allows
for more complete collection-of-urine activity.
4.6 Extrapolation of Results
Using one species to estimate drug behavior in a different species is

an area of great uncertainty. The case below was taken from Princi-
ples of Animal Extrapolation (Calabrese 1983).
If we just consider the size of an elephant, we can easily make

some serious mistakes. A few years ago the journal Science pub-
lished a note which described the reaction of a male elephant to
LSD. The investigators wanted to study the peculiar condition of
the male elephant known as "musth". A male elephant in "musth"
is violent and uncontrollable... Shortly after the publication of
this note, a letter to the editor of Science described the calcula-
tion of this dose of LSD as "an elephantine fallacy" (Harwood,
Science 1963). The authors had calculated the dose, based on the
amount that puts a cat into a rage, and had multiplied it by the
weight until they arrived at 297 mg of LSD to be given to the ele-
phant. The long description of what happened can be shortened
by saying that after the injection of 297 mg (enough for 1,500 hu-
man "trips", for a single human dose is about 0.2 mg), the ele-
phant immediately started trumpeting and running around, then he
stopped and swayed; 5 min after the injection he collapsed, went
into convulsions, defecated, and died.
The authors concluded that elephants are peculiarly sensItlve to

LSD. It is far more likely that the investigators' chosen extrapolation
method was flawed. The choice of a particular extrapolation method
is really an area of freedom for the researcher. There are several ap-
proaches to characterizing interspecies differences, and no one meth-
od has yet clearly distinguished itself as the gold standard. One can
(1) use the organ data directly, i.e., assume that the percentage at
time t in the animal equals the percentage at time t in the human,
(2) use similarity ratios to extrapolate the data, where similarity ra-
tios are the ratio of body weight, metabolic rate, or surface area for
the two species raised to some power, (3) extrapolate based on tis-
sue concentrations as a function of body weight, or (4) try to devel-
op a physiologically based pharmacokinetic (PBPK) model using the
animal data with a compartmental model assumed to be appropriate
86 M.G. Stabin
for humans. Some success can be found for all of these methods for
selected compounds, but the point to remember is that a mouse is
not a little man, nor a rabbit or even a monkey. Even if a reasonable
extrapolation model is chosen and systematically applied, the dose
estimates are always preliminary and must be confirmed with human
studies. Usually a well-done animal study will identify the critical
organs and routes of excretion and give reasonably good estimates
of absorbed dose to be expected in humans, but the numbers should
always be considered in light of their inherent limitations.
The pitfalls in the use of animal data for extrapolation of human
dose estimates are numerous. Capable researchers doing careful
work have been often surprised at how far predictions have strayed
from observed results. This cannot preclude the use of preclinical
data, of course; it merely emphasizes the caution needed in interpret-
ing animal data. Some surprises are unavoidable; examples include
the poor prediction of cardiac uptake by dogs of F-18 fluorodeoxy-
glucose (FDG) (Christman 1980) and 99ffiTc-Iabeled cations
(Deutsch 1985) and the overprediction of 1-123 N-isopropyl-p-
[1231]iodoamphetamine (IMP) uptake in the eyes by baboons (Hol-
man et al. 1983). Sparks and Aydogan (1999) evaluated the perfor-
mance of some commonly used data extrapolation techniques to pre-
dict residence times in humans using animal data from a number of
radiopharmaceuticals. They were not able to form solid conclusions
about a "best" method to use, but found that a mass-based extrapola-
tion provided no significant improvement over no extrapolation in
predicting human dosimetry, that a time-based extrapolation did pro-
vide a significant improvement over no extrapolation, but that a
mass-and-time-based extrapolation did not provide a further im-
provement. In each approach, however, there were significant under-
or overestimations of organ residence time and dose; again, no clear
method emerged as a preferred method.
4.7 Kinetic Analysis
In general, there are three levels of complexity that a kinetic analy-

sis can take:
1. Direct Integration. One can directly integrate under the actual

measured values by a number of methods. This does not give very
much information about your system, but it does allow you to calcu-
late r rather easily. The simplest method for estimating the area un-
der a curve is by approximating the area by a series of trapezoids.
2. Least Squares Analysis. An alternative to "brute-force" integration

of the data is to attempt to fit curves of a given shape to the data.
The curves are represented by mathematical expressions that can be
directly integrated. The most common approach is to attempt to
characterize a set of data by a series of exponential terms, as many
systems are well represented by this form, and exponential terms are
easy to integrate. In general, the approach is to minimize the sum of
the squared distance of the data points from the fitted curve.
The curve will have the form:
The method looks at the squared difference between each point and
the solution of the fitted curve at that point, and minimizes this
quantity by taking the partial derivative of this expression with re-
spect to each of the unknowns aj and bj and setting it equal to zero.
Once the ideal estimates of aj and bi are obtained, the integral of
A(t) from zero to infinity is simply:
1o
00
A(t)dt
al
bl
a2
= -+-+ ...
b2
If the coefficients aj are in units of activity, this integral represents

cumulated activity. If the coefficients are in units of fraction of in-
jected activity, the integral is exactly r.
3. Compartment Model Analysis. The situation frequently arises that

you either know quite a bit about the biological system under inves-
88 M.G. Stabin
tigation or you would like to know in greater detail how this system
is working. In this case, you can describe the system as a group of
compartments linked through transfer rate coefficients. Solving for T
of the various compartments involves solving a system of coupled
differential equations describing transfer of the tracer between com-
partments and elimination from the system. The SAAM II software
(Foster and Barrett 1999) can greatly facilitate such an analysis, and
even performs iterative adjustment of model parameters to fit gath-
ered data in an organ or complete system.
In any analytical method chosen, it is very important to document
all assumptions and calculations made. In particular, it is essential to
(1) know whether the data collected include or do not include radioac-
tive decay (i.e., whether the half-times, in cases 2 and 3, are "biolog-
ical" or "effective"), and (2) to describe what assumptions were made
regarding extrapolation of the kinetic analysis beyond the last observed
datum. The trapezoidal method is particularly weak in this area. The
normal assumption to make is that only physical decay of the radionu-
clide occurs beyond the last data point. Alternatively, one may look at
the slope of the last two or three points and assume that this slope con-
tinues until zero activity is reached. In the least squares approach, one
may assume that the last exponential term continues to infinite time, or
that it continues only to the last time point, and that physical decay
only occurs after this. An interesting characteristic of compartment
models is that it is completely reasonable to assume that the kinetic
behavior observed during the experiment continues to infinite time.
This is true, as one assumption of the model is that the transfer rates
between compartments are constant over time.
4.8 Application of Dosimetric Models
Once the first "half' of our absorbed dose equation is solved, by

successful completion of a kinetic study, one must apply dose con-
version factors to the resultant time-activity integrals in order to cal-
culate the dose to various organs within the body. To do this, one
must employ some model of the body and its individual organs (as
one cannot measure the actual dose in the structures). The first fairly
realistic model of the full human body was the Fisher-Snyder hetero-
geneous, hennaphrodite, anthropomorphic model of the human body

in the 1970s (Snyder et al. 1978). This model, or "phantom", con-
sisted of spheres, ellipsoids, cones, tori, and subsections of such ob-
jects, combined to approximate the geometry of the body and its in-
ternal structures. For the majority of cases, dose from deposited
electron energy was calculated assuming that all energy was ab-
sorbed where it was created (exceptions included hollow organs with
separate wall and contents sections, and the bone and marrow, in
which "crossfire" between adjacent regions can occur due to the
small dimensions of the regions). For photons, Monte Carlo methods
were developed using a computer code called ALGAM (Warner and
Craig 1968), which created photons at random positions within any
"source" region (organ or tissue assumed to be contaminated with
radioactivity), gave these photons a random orientation in 4p space,
and then followed them through various Compton and photoelectric
interactions (pair productions events were quite rare, as starting ener-
gies did not exceed 4 MeV) until the photon reached a certain criti-
cal low ("cut-off') energy and was assumed to be locally absorbed,
or until it escaped the surface of the body (at which point the prob-
ability of scatter from an air molecule and redirection towards the
body was assumed to be negligibly low).
With repeated sampling of the source, which in this time gener-
ally involved only tens of thousands of trials (histories), a statistical
average behavior of particles originating in this source could be ob-
tained for other regions of the body of interest to radiation dose as-
sessment ("target" regions). This behavior was reported as the frac-
tion of energy emitted in the source that was absorbed in a target
(absorbed fraction, rjJ, above). The original phantom developed was
intended mainly to represent a "Reference Man", as defined by the
ICRP from an extensive collection of medical and other literature,
restricted primarily to European and North American populations.
Both due to the makeup of the nuclear medicine population and the
diversifying worker population, the need for other phantoms arose.
Cristy and Eckennan (1987) of Oak Ridge National Laboratory
(ORNL) developed a series of phantoms representing children of dif-
ferent ages, one of which (15 years old) also served as a model for
the adult female. Absorbed fractions (AFs) for these phantoms were
developed using the ALGAMP code ("P" signifying a "parameter-
90 M.G. Stabin
ized" version of the code - allowing substitution of parameters giv-

ing the radii and positions of the various organs at different ages)
(Ryman et al. 1987). These values were published in an ORNL
document, but never officially adopted in the MIRD or other peer-re-
viewed literature. Nonetheless, they were widely accepted and used
for dose calculations in different age individuals. In 1995, Stabin et
al. (1995) released a set of phantoms representing the adult female
at different stages of gestation, to satisfy the need to calculate radia-
tion doses to the fetus from nuclear medicine and other applications.
These calculations were also done with the ALGAMP code.
4.9 Development of MIRDOSE Codes
Calculation of internal dose requires repetitive calculations that are

quite boring, and prone to error when done by hand. The MIRDOSE
computer program (Stabin 1996) was originally developed to elimi-
nate the tedium of repetitive internal dose calculations (looking up
dose conversion factors from tables and adding contributions from
every source to every target, even if of minor importance) and to
automate the calculation of organ doses in nuclear medicine. The
original code was developed in 1983 on a Tektronix stand-alone
workstation, which had 32 Kb of total memory and storage limited
to 8-inch (20.32 em) soft disks (Watson and Stabin 1984). This ver-
sion (MIRDOSE 1) was never distributed. In its second incarnation
in about 1985, MIRDOSE 2, the code was rewritten to be usable on
the International Business Machines Corporation (IBM) Personal
Computer, which was rapidly growing in popUlarity. The code em-
ployed 59 radionuclides, and had available the AFs from the Cristyl
Eckerman phantom series. Users entered cumulated activities for the
source organs involved, and obtained dose estimates for chosen tar-
get organs, with contributions from all source organs included, and
appropriate "remainder of the body" corrections (Cloutier et al.
1973) applied.
The MIRDOSE 3 code represented the migration of the MIR-
DOSE code from the DOS to the Windows environment. Written in
the Microsoft Visual Basic environment, users could select from a
menu of over 200 radionuclides, whose decay data can be used with
absorbed fractions from any of 10 phantoms (the six CristylEcker-

man pediatric phantoms and the four phantoms representing the non-
pregnant or pregnant adult female). The user entered residence times
(Loevinger et al. 1988) for all source organs involved, and the pro-
gram calculated and applied the appropriate S values to obtain organ
dose estimates (for all available, not selected, target regions). The
program also provided calculation of the ICRP-defined effective
dose equivalent (ICRP 1979) and effective dose (ICRP 1991). The
software thus facilitated dose calculations under different conditions,
in different phantoms, etc. The program has been used to calculate
dose estimates for a variety of nuclear medicine compounds, and its
use is widely cited in the literature.
The evolution of the MIRDOSE software continues today. In or-
der to provide more computer platform independence, Dr. Stabin re-
cently rewrote the MIRDOSE code entirely in the Java language and
incorporated a curve-fitting algorithm for kinetic data developed by
Dr. Sparks (Stabin and Sparks 1999). Java is compiled on each ma-
chine before execution, thus allowing the code to be used on Unix,
MacIntosh, Windows NT, and other platforms. The code was re-
named "OLINDA" (Organ Level INternal Dose Assessment), partly
to distinguish it from the activities of the MIRD Committee (which
had expressed concern that the name MIRDOSE might imply that it
was a product of that committee), and partly to integrate the name
into a new unified system of internal and external dose assessment.
This unified system is deployed on an internet web site for rapid
electronic access (Stabin et al. 2001; www.doseinfo-radar.com).This
site, called the RAdiation Dose Assessment Resource (RADAR) pro-
vides decay data for over 800 radionuclides, absorbed fractions for
all available stylized phantoms and some voxel phantoms, kinetic
data, dose factors (for all phantoms and nuclides), risk information,
and other data via electronic transfer to users worldwide.
4.10 Conclusions
Accurate calculation of internal dose estimates requires a number of

complex and tedious steps that must be done with an attention to de-
tail. The use of animal models to predict human kinetics and dosi-
92 M.G. Stabin
metry is an essential first step in the evaluation of new drug safety,

but involves many uncertainties. The data themselves have inherent
uncertainties. There is no "gold standard" method for extrapolating
animal data to humans. Any chosen method will produce a guess at
the correct answer; a more reliable answer will necessarily be ob-
tained by further studies in humans. Performing a kinetic analysis on
the extrapolated data is also an area where the researcher has several
options. All of them are basically acceptable, if the work is done
carefully and careful documentation of methods and results is pro-
vided. Finally, choice of a physical model for applying the kinetic
data and obtaining doses has been automated through several avail-
able low-cost or free software tools which provide not only automa-
tion of the calculations but standardization of method as well. This
standardization is important not only to the scientific method but
also to the regulatory approval process. The careful application of
all of the above steps by knowledgeable professionals can result in a
reliable evaluation of the radiation dosimetry for a radiopharmaceuti-
cal product.
References
Calabrese EJ (1983) Principles of animal extrapolation. Wiley, New York

Christman DR (1980) Tissue distributions of radiopharmaceuticals with posi-
tron emitters and problems relating them to human studies. Third Interna-
tional Radiopharmaceutical Dosimetry Symposium, Oak Ridge, pp 250-
258
Cloutier R, Watson E, Rohrer R, Smith E (1973) Calculating the radiation
dose to an organ. J Nucl Med 14:53-55
Cristy M, Eckerman KF (1987) Specific absorbed fractions of energy at var-
ious ages from internal photon sources. I Methods, II one year old, III
five year old, IV ten year old, V fifteen year old male and adult female,
VI new-born and VII adult male. ORNLlTM 8381Nl-V7, Oak Ridge
National Laboratory
Deutsch E (1985) Proceedings of the 26th annual meeting, Southeastern
Chapter Society of Nuclear Medicine, pXl
Foster D, Barrett P (1999) Developing and testing integrated multicompart-
ment models to describe a single-input multiple-output study using the
SAAM II software system. In: Proceedings of the Sixth International
Radiopharmaceutical Dosimetry Symposium, Oak Ridge Institute for
Science and Education, pp 577-599
Holman BL, Zimmennan RE, Schapiro JR, Kaplan ML, Jones AG, Hill TC
(1983) Biodistribution and dosimetry of N-isopropyl-p-[123IJ iodoamphe-
tamine in the primate. J Nuel Med 24:922-931
International Commission of Radiation Units and Measurements (1980)
ICRU report 33, radiation quantities and units, ICRU 33. International
Commission on Radiation Units and Measurements
International Commission on Radiological Protection (1979) Limits for intakes
of radionuclides by workers. ICRP Publication 30. Pergamon Press, New York
International Commission on Radiological Protection (1991) 1990 recom-
mendations of the International Commission on Radiological Protection.
ICRP Publication 60. Pergamon Press, New York
Loevinger R, Budinger T, Watson E (1988) MIRD primer for absorbed dose
calculations. Society of Nuelear Medicine, New York
Ryman J, Warner G, Eckennan K (1987) ALGAMP - a Monte Carlo radia-
tion transport code for calculating specific absorbed fractions of energy
from internal or external photon sources. Oak Ridge National Laboratory
Report ORNLffM-8377
Siegel J, Thomas S, Stubbs J, Stabin M, Hays M, Koral K, Robertson J, Ho-
well R, Wessels B, Fisher D, Weber D, Brill A (1999) MIRD Pamphlet
no 16 - techniques for quantitative radiophannaceutical biodistribution
data acquisition and analysis for use in human radiation dose estimates. J
Nuel Med 40:37S-61S
Snyder W, Ford M, Warner G (1978) Estimates of specific absorbed frac-
tions for photon sources unifonnly distributed in various organs of a het-
erogeneous phantom. MIRD Pamphlet No.5, revised. Society of Nuelear
Medicine, New York
Sparks R, Aydogan B (1999) Comparison of the effectiveness of some com-
mon animal data scaling techniques in estimating human radiation dose.
In: Proceedings of the Sixth International Radiopharmaceutical Dosimetry
Symposium, Oak Ridge Institute for Science and Education, pp 705-716
Stabin M (1996) MIRDOSE - the personal computer software for use in in-
ternal dose assessment in nuelear medicine. J Nuel Med 37:538-546
Stabin MG, Sparks RB (1999) MIRDOSE 4 does not exist. J Nuel Med
(Suppl 40):309P
Stabin M, Watson E, Cristy M, Ryman J, Eckerman K, Davis J, Marshall D,
Gehlen K (1995) Mathematical models and specific absorbed fractions of
photon energy in the nonpregnant adult female and at the end of each tri-
mester of pregnancy. ORNL Report ORNLffM-12907
Stabin M, Siegel J, Hunt J, Sparks R, Lipsztein J, Eckennan K (2001) RA-
DAR - the radiation dose assessment resource. An online source of dose
infonnation for nuelear medicine and occupational radiation safety (ab-
stract). J Nuel Med 42:243P
Warner GG, Craig AMJ (1968) ALGAM, a computer program for estimat-
ing internal dose for gamma-ray sources in a man phantom. ORNL-TM-
2250, Oak Ridge National Laboratory
94 M. G. Stabin: Internal Dose Assessment
Watson E, Stabin M (1984) Basic alternative software package for internal

radiation dose calculations. In: Kathren RL, Higby DP, McKinney MA
(eds) Computer applications in health physics. Proceedings of the 17th
Midyear Topical Symposium of the Health Physics Society. Columbia
Chapter, HPS, Richland
5 PET for Drug Development
C. Halldin, B. Gulyas, L. Farde
5.1 Introduction................................. 95
5.2 Development of PET and SPECT Receptor Radioligands .... 98
5.3 Evaluation of Radioligands . . . . . . . . . . . . . . . . . . . . . . . 101
5.4 Measuring and Modelling Drug and Physiological Effects
in the Brain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
5.5 Two Possible Experimental Designs . . . . . . . . . . . . . . . . .. 103
5.6 Indirect Approach: Drug-Ligand Interactions . . . . . . . . . . .. 105
5.7 Direct Approach: Radiolabelling of Drugs in Tracer Doses . .. 106
5.8 Individual Therapy Planning - Eliminating Side Effects and
Establishing the Optimal Individual Dose from Clinical Trials. 107
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 109
5.1 Introduction
The development of new drugs involves an extensive pre-clinical

characterization and safety documentation followed by a time-con-
suming search for appropriate clinical dose levels. Drug develop-
ment is a time-consuming and costly procedure: The molecule-to-
drug-time is about 15 years, whereas an investment of about
500 million to 1 billion euro is required to reach registration of a
new drug (Campbell 1995; Halldin et al. 2001 a) (Fig. 1). The ad-
vent of positron emission tomography (PET) with allied techniques
has resulted in a revolutionary change in this respect, as the applica-
tion of PET in drug development and testing can significantly re-
duce both molecu1e-to-drug time and costs (Maziere et al. 1991; Ma-
ziere and Delforge 1994; Campbell 1995; Halldin et al. 1995,
96 C. Halldin et al.
Fig. 1. The process of drug development (modified after Campbell 1995)
2001 a; Farde 1996; Bums et al. 1999; Fowler et al. 1999; Gibson et
al. 2000; Eckelman 2002).
Molecules labelled with short-lived radionuc1ides have been used
extensively to study transformations and distributions of endogenous
compounds and pharmaceuticals, since the method allows for detec-
tion of very low levels of materials. PET and single photon emission
computed tomography (SPECT) are molecular imaging techniques
that use radiolabelled molecules to image molecular interactions of
biological processes in vivo. They are useful techniques to track
transformed compounds with a common origin without the need for
development of specific analytical methods for each of the constitu-
ents. The development of new methods to measure and display
radioactivity in three dimensions by computerized tomography is in
an advanced stage.
PET has been widely used to visualize and quantify different bio-
chemical processes and parameters such as metabolic processes and
receptor densities and localize them to anatomical structures. The
PET for Drug Development 97
PET technique utilizes radiotracers labelled with relatively short-

lived positron emitting radionuclides such as ISF or ultrashort-lived
radionuclides such as IIC and 150. With this technique, minute
amounts of radiotracers can be used due to the very high specific
radioactivity obtainable by the short-lived radionuclides. SPECT is
used to visualize and measure the relative concentration of radioac-
tivity in tissue after injection of compounds labelled with a relative-
ly short-lived single photon emitting radionuclide such as 1231. Bio-
chemical changes can be determined by PET and SPECT in order to
monitor pathological conditions in living humans even before any
anatomical defects occur. This has a great diagnostic value and gives
new information about disease states and their potential therapeutic
treatments. Radiolabelling of endogenous compounds such as glu-
cose, amino acids and acetate have been used in studies on metabo-
lism and synthesis of various human tissues such as tumour, cardiac
and brain tissue. While both PET and SPECT can detect radiotracer
distribution, the SPECT technique is more readily available than
PET. The development of a new generation of SPECT cameras,
namely coincidence positron imaging using a gamma camera, should
significantly increase PET utilization. PET requires a cyclotron to
generate the radionuclides preferable in close connection to the
radiochemistry laboratory and the PET camera. However, the PET
technique offers several advantages, such as the ability to measure
the concentration of the tracer quantitatively, a greater sensitivity, a
higher spatial resolution and chemically more diverse radiotracers.
In addition, the recently developed combined PET-CT camera is an
excellent contribution, which enables a more anatomically correct
image analysis.
During the past decades, over a hundred neurotransmitters have
been identified in the primate brain, including the human brain.
Most of the currently used drugs for the treatment of psychiatric and
neurologic disorders interact with central neurotransmission. Several
receptor subtypes, transmitter carriers and enzymes have proved to
be useful targets for drug treatment. Molecular biological techniques
have now revealed the existence of several novel receptors for which
little or no prior pharmacological or functional data existed. Due to
the lack of data on the functional significance of these sites, pharma-
cologists are now challenged to find the physiological roles of these
98 c. Halldin et al.
receptors and identify selective agents and possible therapeutic indi-

cations.
For several neuropsychiatric disorders there is a lack of generally
accepted animal models. This is a limitation when exploring the
pharmacological significance of new biochemical targets. For in-
stance, the lack of effect in traditional behavioural pharmacology in
animals does not exclude the prospect of a molecule for the treat-
ment of subjectively reported syndromes, such as anxiety or thought
disorders, in the more complex human brain. There is an increasing
awareness that more efficient and sensitive strategies must be ap-
plied in the search for useful medicines. One such strategy is to find
methods to test more molecules in early exploratory studies in man.
5.2 Development of PET and SPECT

Receptor Radioligands
The initial selection of radioligands for neurotransmitter binding

sites, such as receptors, neuronal uptake systems and vesicular up-
take systems is often guided by data obtained in vitro by using tri-
tiated or iodinated radioligands or by displacing a reference radioli-
gand with the unlabelled molecule (Halldin et al. 2001 b). In vitro
binding normally provides information regarding ligand affinity
(e.g., the dissociation equilibrium constants Kd or K i ), and selectivity
(i.e. the relative affinity to competing binding sites) as well as re-
garding the concentration of binding sites (Bmax). The optimum af-
finity is closely related to the expected Bmax. It is preferable if the
Bmax clearly exceeds the Kd of a ligand, i.e. if a binding site exists
in vivo at nanomolar concentrations, a potentially successful radioli-
gand ideally should have a subnanomolar affinity. Binding affinity is
an important factor that determines the ratio of specific binding to
non-specific binding. The higher the ratio the more sensitive the sig-
nal is likely to be to changes in available binding site concentration,
caused by disease or drug occupancy.
Binding affinity (i.e. the fraction of dissociation rate constant,
koff, and association rate constant, k on ) usually governs the approach
to be taken in the biomathematical modelling of the ligand-receptor
interaction (Halldin et al. 2001 b). If the binding of a radio ligand is
reversible over the time scale of a PET experiment, the equilibrium

approaches towards quantification can be utilized. At the opposite
extreme, irreversible ligands normally demand kinetic modelling,
wherein the transfer of radioligand between pharmacological com-
partments (e.g., plasma, tissue, receptors) is described in terms of
rate constants. This approach requires in most cases the determina-
tion of an input function (i.e., the time course of free radioligand in
plasma), which makes the measurement of radio ligand metabolites
in arterial plasma necessary. Very high binding affinity of a radioli-
gand in combination with a comparatively slow clearance from tis-
sue can restrict its usefulness for PET, as the rate-limiting step of
tracer retention may become the delivery instead of the binding pro-
cess (flow-limited conditions).
A further important criterion for a radioligand is binding selectiv-
ity. Ideally, the affinity of a radioligand should be highest for the
site of interest by more than one order of magnitude. However, lack
of selectivity may be acceptable if non-target sites are separated ana-
tomically from the target binding sites. In the light of advances in
molecular biology and pharmacology, the term selectivity often
needs to be revised. Most neurotransmitter receptors have now been
found to exhibit multiple subtypes, and ligands that were initially
thought to bind to a single class of receptors truly display affinity to-
ward several subtypes.
Another substantial consideration in the development of a new
radioligand is estimation of non-specific binding. This is an essen-
tially non-saturable component of the total tissue uptake of a radioli-
gand, usually attributed to adhesion to proteins and lipids. Non-spe-
cific binding and its clearance in vivo are difficult to predict with
absolute confidence. However, within a class of structurally related
compounds, non-specific interactions with tissue generally increase
with increased lipophilicity. The logarithm of the partition coeffi-
cient between water (or preferably buffer to account for ionization at
physiological pH) and octanol (log P) is often taken as a useful in-
dex for the lipophilicity of a compound in the context of biological
systems. Conversely, some degree of lipid solubility is needed for
good passage over the blood-brain barrier, which is a prerequisite
for satisfactory counting statistics. However, the lipophilic nature of
a molecule might also favour binding to plasma proteins, thus reduc-
ing the available "free fraction" in blood that is capable of diffusing

through membranes. Moreover, lipophilic molecules can be extracted
and metabolized in lung tissue when passing through the lung circu-
lation, which prevents them from reaching their sites of action.
Taken together, it appears that there is an optimal - but rather nar-
row - "window" of lipophilicity for brain radioligands, wherein
brain uptake is high and non-specific binding comparatively weak.
Fig. 2. Suitable PET radioligands for the dopamine and serotonin receptor/
transporter systems
PET measures the regional radioactivity concentration without

being able to distinguish the chemical forms or environments in
which the radioactivity resides. For a clearly interpretable signal it is
therefore necessary that radiolabelled metabolites do not contribute
to specific binding. Thus, radioligands should be preferably resistant
to rapid metabolism over the period of data acquisition. Further-
more, radiolabelled metabolites should not be taken up or retained
in the target area, or both. This requirement may have important
consequences concerning the elaboration of a radiolabelling strategy.
A very important consideration in the context of radiochemistry is
specific radioactivity (SR) of the radioligand. Too low SR may re-
sult in pharmacological effects or toxicity of the radiotracer. More-
over, low SR may saturate the biological system of interest, thus
abolishing the mandatory tracer conditions. For low-density binding
sites, very high SR is essential in order to exclude a substantial oc-
cupation of target sites by unlabelled ligand.
It has to be emphasized that any data extracted from in vitro ex-
periments can only give a rough estimate of the situation to be en-
countered in vivo. Most in vitro assays use homogenized tissue,
which does not reflect the tissue heterogeneity in the intact organ in
vivo. Competition with endogenous ligands may lower the binding
of a radioligand at a given site. It has to be kept in mind that neuro-
transmission systems in the intact body constitute part of a dynamic
and communicating environment and that neural interaction may ac-
tually alter in vivo receptor binding. Some useful PET radioligands
used worldwide for the dopamine and serotonin receptor/transporter
systems are shown in Fig. 2.
5.3 Evaluation of Radioligands
When a potential candidate ligand has been identified and a label-

ling technique developed, some preclinical evaluation, prior to PET
or SPECT in humans, needs to be performed (Halldin et al. 2001 b).
Some useful information may be obtained by studies in rodents.
Typically, the radioligand is injected intravenously into a series of
rodents. These are then sacrificed at known times after injection and
different organs are removed and radioactivity is counted, thus pro-
viding the distribution of the radiotracer in different organs over

time (tissue distribution study). In addition to that, clearance of
radioactivity from plasma and information on the appearance of la-
belled metabolites in plasma can be obtained. Specificity of binding
may be demonstrated by using selective and potent unlabelled com-
pounds that act in a competing way at the site of interest. Small-ani-
mal PET devices are available which promise to simplify radioli-
gand evaluation. However, species differences may sometimes lead
to results which do not extrapolate to humans.
A complementary tool in early radioligand development may be
autoradiography experiments, wherein frozen slices of tissue ob-
tained from the organ of interest (e.g., brain) are mounted on glass
slides and incubated for a given time with a buffered radioligand so-
lution. These sections are exposed to radiation-sensitive film. Auto-
radiography may provide information if a radioligand is suitable for
PET, especially regarding affinity, selectivity and non-specific bind-
ing. An advantage compared to in vitro homogenate binding assays
is the use of intact tissue, which provides information in an anatomi-
cal manner. However, no data regarding the in vivo pharmacoki-
netics of the ligand can be deduced from such an experiment. This
lack of information can be compensated for by ex vivo autoradiogra-
phy approaches, where the analysis is done in vitro after in vivo ad-
ministration of the radioligand into small animals. Traditionally,
autoradiography experiments are performed with 3H_ or 125I-labelled
compounds, but the use of fJ-sensitive film makes the procedure also
suitable for molecules labelled with positron emitters, even though
this lowers the achievable spatial resolution. A utile variant of this
method is whole hemisphere autoradiography of the postmortem hu-
man brain.
The next step in radioligand development is normally PET in ani-
mals. For brain imaging studies non-human primates - such as cyno-
molgus monkey (Macaca fascicularis) - are preferred. Analysis of
labelled metabolites from venous blood samples provides useful in-
formation regarding clearance and metabolic pathways. Administra-
tion of potent and selective competing compounds prior to radioli-
gand injection (pretreatment) or during the time course of the PET
experiment (displacement) can demonstrate the specificity and rever-
sibility of radioligand binding. It should be noted, however, that spe-
cies differences may be encountered and lead to different results be-

tween animals and human subjects.
5.4 Measuring and Modelling Drug

and Physiological Effects in the Brain
The great advantage of the PET technique is that it is capable of ob-

taining absolute measurements of regional radioactivity concentra-
tions which, in tum, with the help of appropriate tracer kinetic mod-
els can be transformed into quantitative parametric maps of various
biochemical, physiological or neurotransmitter/receptor system-re-
lated parameters. The quantitative measurements with PET require
the determination of tissue and blood/plasma radioactivity concentra-
tions and the a priori knowledge of a number of experimental pa-
rameters, including basic features related to the tracer, the biochem-
ical and physiological characteristics and metabolic stability of the
tracer/ligand, and the those of the scanner. In the next step, a multi-
compartmental model describing the distribution and metabolism of
the ligand in the brain and, eventually, in other body compartments,
is being developed, tested and validated. With the help of appropri-
ate tracer kinetic models, the requested biological variables, e.g., re-
ceptor occupancy data, can be described in quantitative terms in pre-
cise anatomical context in the organ covered by the PET scan.
5.5 Two Possible Experimental Designs
After the first PET scanners proved to be uniquely useful in medical

diagnostics and basic biological research, PET entered the armoury
of drug research and development. During the past two decades the
accumulated experience with PET has clearly shown that the appli-
cation of the technique is especially useful and promising in the
field of neuropsychopharmacology (Halldin et al. 200la) (Table 1).
There are two possible experimental designs widely used in neu-
ropsychopharmacological drug development with PET (Halldin et al.
2001 a):
Table 1. Applications of PET during drug development
Pharmacokinetics (movement of drugs):

1. To confirm brain distribution (i.e. passage across the blood-brain
barrier)
2. To confirm that the drug binds to central neuroreceptors (validation
of principle)
3. To identify the relationship between dose, plasma concentration
and central receptor occupancy
4. To provide additional information on absorption, bioavailability,
distribution and elimination
Pharmacodynamics (therapeutic and side effects):
1. Correlation between receptor binding and therapeutic effects and
side effects
2. Correlation between regional binding and regional effects on physiolo-
gical parameters (metabolism, blood flow, etc.)
3. Functional measurements between and after treatment with the drug
Drug testing:
1. Validation of animal models for human conditions
2. Measuring species differences (tissue metabolism, receptor binding)
3. Comparison between the effects of various drug molecules
Therapy planning:
1. Individual therapy planning
2. Optimizing therapeutic effects
3. Minimizing side effects
1. The most frequent approach is to study how an unlabelled drug

inhibits specific binding of a well-characterized selective PET
radioligand. In this case, the unlabelled drug is entered into com-
petition with a labelled ligand for occupying a receptor system.
The amount of the unlabelled drug excessively exceeds that of
the labelled drug. The unlabelled drug is given in pharmacologi-
cal dose (mg range) whereas the labelled ligand is given in tracer
dose; the two doses being different in at least three or four orders
of magnitude. The reason behind this is that we would like to
block a given receptor system as completely as possible by the
unlabelled drug. This can be achieved either by giving the unla-
belled drug before the administration of the labelled drug (pre-
treatment) or somewhat after it (displacement).
2. The alternative direct approach is to radiolabel a new potential

drug and to trace its uptake, anatomical distribution and binding
in brain.
In addition to these "main" uses of PET in neuropsychopharmacolo-

gical drug development, the technique of assessing the effects of a
neuropsychopharmacological drug on metabolic or blood flow pa-
rameters can also be used, thereby assessing its therapeutic efficacy
in relation to vital biochemical-physiological parameters. Further-
more, with PET the sites of pharmacological actions and physiologi-
cal actions can be identified, providing clues for interpreting drug ef-
fects. The technique in combination with other measurements can
also yield information about metabolic routes of drug metabolism.
5.6 Indirect Approach: Drug-Ligand Interactions
A molecule may be an excellent drug but not suitable for PET visua-
lization of binding to the target receptor. The binding affinity may
be too low or the molecule may not be selective or lack the chemi-
cal structure needed for rapid labelling with a positron emitting
radionuclide. In such cases, an indirect approach with PET is to
measure how the unlabelled drug inhibits ("blocks") selective radio-
ligand binding in the eNS, i.e., how receptor occupancy is modified
by the drug. A major advantage with this indirect approach is that
drug binding can be examined at clinically relevant dose levels. Re-
ceptor occupancy is defined as the fraction (%) of the receptor popu-
lation which is occupied by the unlabelled drug. So far, the PET
measurements of receptor occupancy have most extensively been ap-
plied to the study of antipsychotic drug binding. Experimental stud-
ies have shown that all neuroleptic drugs have affinity for the D2-
and D3-dopamine receptor subtype.
The requirement of high affinity is illustrated by PET studies
with raclopride and remoxipride, two drugs developed for the treat-
ment of schizophrenia (Farde and von Bahr 1990) (Fig. 3). The two
dopamine antagonists are close analogues with affinity for D2- and
D3-receptors and negligible affinity for other central receptors. The
major difference between the two drugs is that raclopride has about
Fig. 3. Distribution of radioactivity in a horizontal brain section through the

caudate-putamen level b~ PET. The tracers are the dopamine D2 antagonists
[llClremoxipride and [ lC]rac!opride. Baseline images are shown on the
left. Pretreatment was made with pharmacological doses of remoxipride
(right)
lOO-fold higher affinity than remoxipride. A consequence is that

[IlC]raclopride provides a good signal for PET visualization of spe-
e
cific binding, whereas the IC]remoxipride binding is to low to be
differentiated from the background of free and non-specific binding.
5.7 Direct Approach: Radiolabelling of Drugs

in Tracer Doses
Before the advent of functional neuroimaging techniques, it has for

obvious reasons not been possible to pursue traditional invasive
methods to measure drug concentrations directly in the human brain.
Attempts have been made to use drug concentration in CSF as an es-
timate for the concentration in brain. The use of imaging techniques
to measure drug concentrations directly in brain is therefore a major
advancement in clinical neuropsychopharmacology. PET can be used
for quantitative "Galenic studies", as it can measure and localize the

anatomical distribution of radiolabelled drug molecules in the human
brain.
The PET technique is sensitive enough for determinations of con-
centrations as low as in the subpicomolar range (Halldin et al.
2001 a). Effective radiochemical labelling provides a drug which is
labelled to a high specific radioactivity, i.e., with a high ratio of radio-
labelled to unlabelled drug molecules. A consequence is that i.v. injec-
tion of less than a microgram of the radiolabelled drug is sufficient for
a PET study in man. The concept "tracer dose" is often used to empha-
size the low mass which does not induce drug effects.
The use of tracer doses has advantages in initial drug develop-
ment. For drug administration to man, approval can be obtained with
less extensive safety and toxicology documentation than that which
is required for the much higher doses which induce clinical effects.
In this emerging field there are as yet no international guidelines for
the preclinical documentation required. A common view applied by
several national drug agencies is to approve PET studies with new
radiolabelled drugs if the acute toxicity has been examined in a ro-
dent and if a genotoxicity test is negative. Due to this fact, it is now
possible to use PET for examination of a large number of new mole-
cules directly in man.
5.8 Individual Therapy Planning -

Eliminating Side Effects and Establishing
the Optimal Individual Dose from Clinical Trials
Individual therapy planning eliminates side effects and is used to es-

tablish the optimal individual dose from clinical trials. PET findings
of high D2-dopamine receptor occupancy in the basal ganglia of pa-
tients responding to antipsychotic drug treatment has provided addi-
tional support for the dopamine hypothesis of antipsychotic drug ac-
tion. A consistent finding has been that chemically distinct classical
neuroleptics all induce a 70-90% central D2-receptor occupancy
(Farde et al. 1992).
Quantitative relationships can thus be examined among complex
therapeutic effects, brain biochemistry, occupancy and dose inter-
Fig. 4. The suggested distinct threshold for antipsychotic effects and extra-
pyramidal syndromes as induced by classical antipsychotic drugs. Due to the
hyperbolic relationship between D2 receptor occupancy and antipsychotic
drug dose (or plasma concentration) there is a narrow interval for optimal
therapeutic treatment
vals. An example is a PET study in neuroleptic drug-treated patients

who suffered from acute extrapyramidal syndromes (EPS) such as
parkinsonism, dystonia and restlessness. It was demonstrated that the
patients with EPS had higher D2-occupancy (above 80%) than those
without side effects. This finding demonstrates a relationship be-
tween molecular events and a pharmacological effect in the living
human brain. Patients with good response to antipsychotic drug
treatment but no EPS had lower D2-occupancy (70-80%). Based on
these observations, a PET-defined dose and occupancy interval was
suggested as optimal for clinical antipsychotic drug treatment
(Fig. 4).
References
Bums HD, Hamill TG, Eng W, Francis B, Fioravanti C, Gibson RE (1999)

Positron emission tomography neuroreceptor imaging as a tool in drug
discovery, research and development. CUff Opin Chern Bioi 3:388-394
Campbell B (1995) Drug development and positron emission tomography.
In: Comar D (ed) PET for drug development and evaluation. Kluwer
Academic Publishers, Dordrecht, pp 1-24
Eckelman WC (2002) Accelerating drug discovery and development through
in vivo imaging. Nucl Med Bioi 29:777-782
Farde L (1996) The advantage of using positron emission tomography in
drug research. Trends Neurosci 19:211-214
Farde L, von Bahr C (1990) Distribution of remoxipride to the human brain
and central D2-dopamine receptor binding examined by PET. Acta Psy-
chiatr Scand 82:67-71
Farde L, Nordstrom AL, Wiesel FA, Pauli S, Halldin C, Sedvall G (1992)
Positron emission tomographic analysis of central Dl-dopamine and D2-
dopamine receptor occupancy in patients treated with classical neurolep-
tics and clozapine - Relation to extrapyramidal side effects. Arch Gen
Psychiatry 49:538-544
Fowler JS, Volkow N, Wang G-J, Ding Y-S, Dewey S (1999) PET and drug
research and development. J Nucl Med 40:1142-1163
Gibson RE, Bums HD, Hamill TG, Eng W, Francis B, Ryan C (2000) Non-
invasive radiotracer imaging as a tool for drug development. CUff Pharm
Des 6:973-989
Halldin C, Swahn C-G, Farde L, Sedvall G (1955) Radioligand disposition
and metabolism - key information in early drug development. In: Comar
D (ed) PET for drug development and evaluation. Kluwer Academic Pub-
lishers, Dordrecht, pp 55-65
Halldin C, Gulyas B, Farde L (2001 a) PET studies with carbon-11 radioli-
gands in neuropsychopharmacological drug development. CUff Pharm
Des 7:1907-1929
Halldin C, Gulyas B, Langer 0, Farde L (2001 b) Brain radioligands - state
of the art and new trends. Q J Nucl Med 45:139-152
Maziere B, Delforge J (1994) Positron emission tomography and pharmaco-
kinetics. In: Welling PG, Balant LP (eds) Pharmacokinetics of drugs.
Springer, Berlin Heidelberg New York, pp 455-480
Maziere B, Maziere M, Delforge J, Syrota A (1991) Contribution of posi-
tron emission tomography to pharmacokinetic studies. In: Rescigno A,
Thakur AK (eds) New trends in pharmacokinetics. Plenum Press, New
York, pp 169-187
6 Future Directions in Molecular Imaging
U. Haberkorn
6.1 Antisense Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 111

6.2 Measurement of Gene Transfer During Gene Therapy
with Suicide Genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
6.3 Measurement of Gene Transfer - In Vivo Reporter Genes . . .. 118
6.4 Protein-Protein Interaction . . . . . . . . . . . . . . . . . . . . . . .. 122
6.5 Design of New Biomolecules for Treatment and Monitoring
of New Therapeutic Modalities . . . . . . . . . . . . . . . . . . . .. 124
6.6 Outlook.................................... 127
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 128
After the identification of new genes, functional information is re-

quired to investigate the role of these genes in living organisms.
This can be done by analysis of gene expression, protein-protein in-
teraction, or biodistribution of new molecules and may result in new
diagnostic and therapeutic procedures which include visualization of
and interference with gene transcription and the development of new
biomolecules to be used for diagnosis and treatment.
6.1 Antisense Imaging
Antisense RNA and DNA techniques were originally developed to

modulate the expression of specific genes. These techniques origi-
nated from studies in bacteria, demonstrating that these organisms
112 U. Haberkorn
are able to regulate gene replication and expression by the produc-

tion of small complementary RNA molecules in an opposite (anti-
sense) direction. Base pairing between the oligonucleotide and the
corresponding target mRNA leads to highly specific binding and
specific interaction with protein synthesis. Thereafter, several labora-
tories showed that synthetic oligonucleotides complementary to
mRNA sequences could downregulate the translation of various on-
cogenes in cells (Zamecnik and Stephenson 1978; Mukhopadhyay et
al. 1991).
However, besides their use as therapeutics for specific interaction
with RNA processing, oligonucleotides have been proposed for diag-
nostic imaging and treatment of tumors. Assuming a total human
gene count between 30,000 and 35,000, calculations which take into
account alternative po1yadenylation and alternative splicing result in
a mRNA number between 46,000 and 85,000 (Claverie 2001). It is
expected that an oligonucleotide with more than 12 (l2-mer) nucleo-
bases represents a unique sequence in the whole genome (Woolf et
al. 1992). Since these short oligonucleotides can easily be produced,
antisense imaging using radiolabeled oligonucleotides offers a high
number of new specific tracers. Prerequisites for the use of radiola-
beled antisense oligonucleotides are ease of synthesis, stability in
vivo, uptake into the cell, accumulation of the oligonucleotide inside
the cell, interaction with the target structure, and minimal nonspecif-
ic interaction with other macromolecules. For the stability of radiola-
beled antisense molecules, nuclease resistance of the oligonucleo-
tide, stability of the oligo-linker complex and a stable binding of the
radionuclide to the complex are required. In this respect, modifica-
tions of the phosphodiester backbone such as phosphorothioates,
methylphosphonates, peptide nucleic acids or gapmers (mixed back-
bone oligonucleotides) result in at least a partial loss in cleavage by
RNAses.
Evidence has been presented of receptor-coupled endocytosis as
the low capacity mechanism by which oligonucleotides enter cells
(Loke et al. 1989; Iversen et al. 1992). Subcellular fractionation ex-
periments showed a sequestration of the oligonucleotides in the nu-
clei and the mitochondria of cervix carcinoma (HeLa) cells (Iversen
et al. 1992). This phenomenon of fractionation, problems with in
vivo stability of the oligonucleotides, as well as the stability of the
Future Directions in Molecular Imaging 113
hybrid oligo-RNA structures may prevent successful imaging of

gene expression. Furthermore, binding to other polyanions, such as
heparin, based on charge interaction result in unspecific signals.
However, successful antisense imaging has been reported in several
studies: accumulation of I I lIn-labeled c-myc antisense probes with a
phosphorothioate-backbone occurred in mice bearing c-myc overex-
pressing mammary tumors (Dewanjee et al. 1994). Imaging was also
possible with a transforming growth factor-a antisense oligonucleo-
tide, an antisense phosphorothioate oligodeoxynucleotide for the
mRNA of glial fibrillary acidic protein, and a 125 I-labeled antisense
peptide nucleic acid targeted to the initiation codon of the luciferase
mRNA in rat glioma cells permanently transfected with the luciferase
gene (Urbain et al. 1995; Cammilleri et al. 1996; Kobori et al. 1999;
Shi et al. 2000). In addition, 90Y-labeled phosphorothioate antisense
oligonucleotides may be applied as targeted radionuclide therapeutic
agents for malignant tumors (Watanabe et al. 1999).
However, data obtained from messenger RNA (mRNA) profiling do
not faithfully represent the proteome because the mRNA content
seems to be a poor indicator of the corresponding protein levels (An-
derson and Seilhamer 1997; Futcher et al. 1999; Gygi et al. 1999). Di-
rect comparison of mRNA and protein levels in mammalian cells
either for several genes in one tissue or for one gene product in many
cell types reveals only poor correlations with up to 30-fold variations.
This might lead to misinterpretation of mRNA profiling results.
Furthermore, mRNA is labile, leading to spontaneous chemical degra-
dation as well as to degradation by enzymes which may be dependent
on the specific sequence and result in nonuniform degradation of RNA.
This phenomenon introduces quantitative biases that are dependent on
the time after the onset of tissue stress or death. In contrast, proteins
are generally more stable, and exhibit slower turnover rates in most
tissues. A substantial fraction of interesting intracellular events is lo-
cated at the protein level, for example operating primarily through
phosphorylation/dephosphorylation and the migration of proteins. In
addition, proteolytic modifications of membrane-bound precursors ap-
pear to regulate the release of a large series of extracellular signals
such as angiotensin or tumor necrosis factor.
Since protein levels often do not reflect mRNA levels, antisense
imaging may not be a generally applicable approach for a clinically
114 U. Haberkorn
useful description of biological properties of tissues. Expression pro-

filing data would be more useful, if mRNA samples could be en-
riched for transcripts that are being translated (Pradet-Balade et al.
2001). This can be achieved by fractionation of cytoplasmic extracts
in sucrose gradients, which leads to the separation of free mRNPs
(ribonucleoprotein particles) from mRNAs in ribosomal preinitiation
complexes and from mRNAs loaded with ribosomes (polysomes).
Since only the polysomes represent actively translated transcripts,
this fraction should be directly correlated with de novo synthesized
proteins. Polysome imaging with nuclear medicine procedures has
not been tried to date or even may be not possible. Therefore, anti-
sense imaging for the determination of transcription by hybridization
of the labeled antisense probe to the target mRNA makes sense in
cases where RNA and protein content are highly correlated. Success-
ful imaging was possible in cases where the expression of the pro-
tein was proven or the gene of interest was introduced by an expres-
sion vector (Dewanjee et al. 1994; Urbain et al. 1995; Cammilleri et
al. 1996; Kobori et al. 1999; Shi et al. 2000). If no correlation be-
tween mRNA and protein exists, the diagnostic use of antisense
imaging seems questionable. Therapeutic applications may use tri-
plex oligonucleotides with therapeutic isotopes such as Auger elec-
tron emitters, which can be brought near to specific DNA sequences
to induce DNA strand breaks at selected loci.
6.2 Measurement of Gene Transfer During Gene Therapy

with Suicide Genes
From basic science we may expect a better understanding of the

mechanisms of carcinogenesis, tumor progression, and patients' im-
mune response. Furthermore, the characterization of tumor cell-spe-
cific properties allows the development of new treatment modalities
such as gene therapy. Some of these new approaches include the
transfer of foreign genes into normal or tumor tissue:
1. Protection of normal tissues (such as bone marrow), which are

normally targets for cytotoxic drugs. This may be achieved by the
transfer of the gene for the drug efflux pump glycoprotein p.
2. Improvement of the host antitumor response by increasing the

antitumor activity of tumor-infiltrating immune-competent cells or
by modifying the tumor cells to enhance their immunogenicity.
This approach relies on the introduction of genes which are re-
sponsible for the production of foreign surface antigens or of
genes which lead to local production and secretion of cytokines.
3. Reversion of the malignant phenotype either by suppression of
oncogene expression or by introduction of normal tumor-suppres-
sor genes. The inactivation of oncoproteins may be performed by
introduction of genes for intracellular antibodies (intrabodies)
against these oncogenes or by the use of antisense oligonucleo-
tides and ribozymes.
4. Direct killing of tumor cells by the transfer of cytotoxic or pro-
drug-activating genes (Haberkorn and Altmann 2001).
For the clinical application of these new treatment modalities nonin-

vasive tools are needed to evaluate the efficiency of gene transfer in
terms of gene transcription. In that respect, nuclear medicine proce-
dures offer a high sensitivity in the picomolar range. Labeling of
substrates with radioactive isotopes and administration of very low
amounts of these tracers allows the assessment of biochemical or
physiological processes without any interference with the phenom-
ena to be studied.
Expression of the herpes simplex virus thymidine kinase (HSVtk)
has been studied in several tumor models after viral as well as non-
viral transfer of the gene. The principle of in vivo HSVtk imaging
was first demonstrated by Saito et al. for the visualization of HSV
encephalitis (Saito et al. 1982). In all studies, enhanced uptake of
specific substrates (Fig. 1) such as 5-iodo-2' -fluoro-2'deoxy-l-ji'-o-
arabinofuranosyluracil (FIAU), 5-fluoro-2' -deoxycytidine (FCdR), 5-
fluoro-l-(2' -deoxy-fluoro-ji'-o-ribofuranosyl)uracil (FFUdR), ganci-
clovir (GCV), 8-e sF]fluoroganciclovir (PGCV), 9-(4-[ 18F]-fluoro-3-
hydroxymethylbutyl)-guanine ([ 18F]FHBG), and 9-[(3- 18F-fluoro-l-
hydroxy-2-propoxy)methyl]-guanine ([18 F]_FHPG) was seen in ge-
netically modified tumor cells in vitro and in vivo (Monclus et al.
1995; Tjuvajev et al. 1995, 1998; Haberkorn et al. 1997a, 1998b;
Morin et al. 1997; Germann et al. 1998; Gambhir et al. 1999; Alaud-
din et al. 1999; Wiebe et al. 1999; de Vries et al. 2000; Haubner et
116 U. Haberkorn
35
30 0-_-------0
0
/
25
+='
~ /
a>
-'" /
20
/
C1l
15..
::J
.S;
~
/
15
U
0
/
·0
c
C1l
/
10
CD /
5
? • MCF7HSVlklMCF7
--0-
~ MH3924AHSVlklMH3924A
0
0 10 20 30 40 50
incubation time (h)

Fig. 1. Ganciclovir uptake ratio between genetically modified Morris hepato-
ma cells (MH3924A) and human mammary carcinoma cells (HeLa) and
their corresponding wild-type cells as a function of time. Mean values and
standard deviations (n = 3)
al. 2000; Hospers et al. 2000; Hustinx et aL 2001). Furthermore,

GCV, FFUdR, and the PIAU uptake were highly correlated to the
percentage of HSVtk-expressing cells and to the growth inhibition as
measured in bystander experiments (Tjuvajev et al. 1995; Haberkorn
et al. 1997 a; Germann et al. 1998). In rats infected with adenovirus
particles there was also a significant positive correlation between the
percentage of injected dose of 8- [18 F]fluoroganciclovir FGCV re-
tained per gram of liver and the levels of hepatic HSVtk expression
(Gambhir et al. 1999).
To elucidate the transport mechanism of the specific HSVtk sub-
strate, GCV inhibition/competition experiments were performed in
rat hepatoma and human mammary carcinoma cells. The nucleoside
transport in mammalian cells is known to be heterogeneous with two
classes of nucleoside transporters: the equilibrative, facilitated diffu-
sion systems and the concentrative, sodium-dependent systems. In

these experiments competition for all concentrative nucleoside trans-
port systems and inhibition of the GCV transport by the equilibra-
tive transport systems was observed, whereas the pyrimidine nucleo-
base system showed no contribution to the GCV uptake (Haberkorn
et al. 1997 a, 1998 b). In human erythrocytes, acyclovir has been
shown to be transported mainly by the purine nucleobase carrier
(Mahony et al. 1988). Due to a hydroxymethyl group on its side
chain, GCV has a stronger similarity to nucleosides and, therefore,
may be transported also by a nucleoside transporter. Moreover, the
3' hydroxyl moiety of nucleosides was shown to be important for
their interaction with the nucleoside transporter (Gati et al. 1984).
In rat hepatoma cells as well as in human mammary carcinoma
cells, the GCV uptake was shown to be much lower than the thymi-
dine uptake (Haberkorn et al. 1997 a, 1998 b). Therefore, in addition
to low infection efficiency of the current viral delivery systems, slow
transport of the substrate, and also its slow conversion into the phos-
phorylated metabolite, is limiting for the therapeutic success of the
HSVtklGCV system. Cotransfection with nucleoside transporters or
the use of other substrates for HSVtk with higher affinities for nu-
cleoside transport and phosphorylation by HSVtk may improve ther-
apy outcome.
In human glioblastoma cells, the effects of cytosine deaminase
(CD) gene transfer were evaluated. When exposed to 3H-fluorocyto-
sine (5-FC) these cells produced 3H-5-FU, whereas in the control
cells only 3H-5-FC was detected (Haberkorn et al. 1996). Moreover,
significant amounts of 5-FU were found in the medium of cultured
cells, which may account for the bystander effect observed in pre-
vious experiments. However, uptake studies revealed a moderate and
nonsaturable accumulation of radioactivity in the tumor cells and
lack of inhibition by hypoxanthine or uracil, suggesting that 5-FC
enters the cells only via diffusion. Although a significant difference
in 5-FC uptake was seen between CD-positive cells and controls
after 48 h incubation, no difference was observed after 2 h incuba-
tion. Furthermore, a rapid efflux could be demonstrated. Therefore,
5-FC transport may be a limiting factor for this therapeutic proce-
dure, and quantitation with positron emission tomography (PET) has
to rely rather on dynamic studies and modeling, including HPLC
118 U. Haberkorn
analysis of the plasma, than on nonmodeling approaches (Haberkorn

et al. 1996). To evaluate the 5-FC uptake in vivo, a rat prostate ade-
nocarcinoma cell line was transfected with a retroviral vector bear-
ing the Escherichia coli CD gene. The cells were found to be sensi-
tive to 5-FC exposure, but lost this sensitivity with time. This may
be due to inactivation of the viral promoter (cytomegalovirus, CMV)
used in this vector. In vivo studies with PET and 18FC showed no
preferential accumulation of the tracer in CD-expressing tumors,
although HPLC analysis revealed a production of 5-fluorouracil
which was detectable in tumor lysates as well as in the blood of the
animals (Haberkorn and Altmann 2001).
A functional analysis of bacterial CD and yeast CD expressed in
COS-I cells showed that both recombinant enzymes utilized cyto-
sine with equal efficacy, but 5-FC was an extremely poor substrate
for the bacterial CD, with an apparent catalytic efficiency 280-fold
lower than that observed for the yeast CD (Hamstra 1999). Further-
more, after retroviral infection of tumor cell lines with the different
genes the IC 50 of 5-FC was 30-fold lower in yeast CD-infected cells
than in cells with expression of the bacterial CD gene. In subcuta-
neous human colorectal carcinoma xenografts in nude mice, in vivo
magnetic resonance spectroscopy was done to measure yeast CD
transgene expression in genetically modified tumors by direct detec-
tion of CD-catalyzed conversion of 5-fluorocytosine to 5-fluorouracil
(Stegman et al. 1999). A three-compartment model revealed a first-
order kinetics, suggesting that the yeast CD was not saturated in
vivo in the presence of measured intratumoral 5-FC concentrations
above the in vitro determined affinity (Km) values.
6.3 Measurement of Gene Transfer -

In Vivo Reporter Genes
In vivo reporter genes are genes which can be visualized noninva-

sively using radiolabeled molecules. In this respect, genes encoding
for enzymes, receptors, antigens, and transporters have been used.
Enzyme activity can be assessed by the accumulation of the metabo-
lites of radiolabeled specific substrates, receptors by the binding and
internalization of ligands, antigens by binding of antibodies, and
transporters by the uptake of their substrates. Since expression of the

HSVtk gene leads to phosphorylation of specific substrates and to
the accumulation of the resulting negatively charged metabolite, this
gene can be used as an in vivo reporter gene. A general problem
with this gene is the fact that the affinity of these specific substrates
for the nucleoside transport systems as well as for the enzyme is
rather low, which may be a limiting factor for cellular accumulation.
Therefore, at present the ideal tracer for HSVtk imaging has not
been identified and more efforts have to be done to synthesize radio-
labeled compounds with improved biochemical properties. In order
to improve the detection of low levels of PET reporter gene expres-
sion a mutant herpes simplex virus type 1 thymidine kinase (HSVI-
sr39tk) has been used as an in vivo reporter gene for PET (Gambhir
et al. 2000). Successful transfer of this mutant gene resulted in en-
hanced uptake of the specific substrates [8- 3H]penciclovir, and 8-
[18F]fluoropenciclovir in C6 rat glioma cells with a twofold increase
in accumulation compared with wild-type HSVtk-expressing tumor
cells.
Two receptor genes have been used to visualize genetically trans-
formed tumors. The dopamine D2 receptor gene is an endogenous
gene which is not likely to invoke an immune response. Further-
more, the corresponding tracer 3-(2' _[18F]-fluoroethyl)spiperone
(FESP) rapidly crosses the blood-brain barrier, can be produced at a
highly specific activity, and is currently used in patients. The tracer
uptake in nude mice after transfection with an adenoviral-directed
hepatic gene delivery system and also in transplanted stable tumor
cells was proportional to in vitro data of hepatic FESP accumula-
tion, dopamine receptor ligand binding, and the D2 receptor mRNA
(MacLaren et al. 1999). In addition, tumors modified to express the
D2 receptor retained significantly more FESP than wild-type tumors.
In modified non-small cell lung cell lines expressing the human
type 2 somatostatin receptor and transplanted in nude mice, a five-
to tenfold higher accumulation of a 99mTc_ or 188Re-Iabeled soma-
tostatin-avid peptide was obtained (Zinn et al. 2000).
The low expression of tumor-associated antigens on target cells
for radioimmunotherapy may be encountered by the transfer of the
specific gene. Therefore, the gene for the human carcinoembryonic
antigen (CEA) was transfected in a human glioma cell line resulting
120 U. Haberkorn
in high levels of CEA expression (Raben et al. 1996). In these modi-

fied tumor cells, high binding of a 1311-labeled CEA antibody was
observed in cell culture experiments as well as by scintigraphic
imaging.
Transchelation of oxotechnetate to a polypeptide motif from a
biocompatible complex with a higher dissociation constant than that
of a diglycylcysteine (GGC) complex has been used to obtain bind-
ing of oxotechnetate to synthetic peptides and recombinant proteins
like a modified green fluorescence protein (Bogdanov et al. 1998).
In these experiments, intramuscular injection of synthetic peptides
bearing a diglycylcysteine motif resulted in accumulation of radioac-
tivity after application of 99mTc-glucoheptonate. Peptides with three
metal-binding GGC motifs showed a threefold higher accumulation
as compared to the controls. This principle can also be applied to re-
combinant proteins which appear at the plasma membrane (Simono-
va et al. 1999).
The gene of the human sodium iodide symporter (hNIS) has been
transferred in a variety of tumor models (Shimura et al. 1997; Man-
dell et al. 1999; Carlin et al. 2000; Cho et al. 2000; Nakarnoto et al.
2000; Smit et al. 2000, 2002; Spitzweg et al. 2000, 2001; Haberkorn
et al. 2001 a, 2003; Boland et al. 2002; La Perle et al. 2002; Sieger
et al. 2003). The corresponding protein seems promising because,
besides iodine, it accepts pertechnetate, which is commonly avail-
able as a substrate. Transfer of the hNIS gene into tumor cells
caused a significant increase in iodide uptake by a factor of 84 to
235. Animal studies with wild-type and hNIS-expressing tumors in
rats showed a maximum uptake after I h and a continuous disap-
pearance of the radioactivity out of the body as well as of the hNIS-
expressing tumors (Fig. 2; Haberkorn et al. 2001 a, 2003; Sieger et
al. 2003). Although the NIS activity is asymmetrical, favoring iodide
influx, there is obviously an efflux activity with the consequence
that in cells that do not organify iodide the concentration of intracel-
lular iodide will drop proportionally to the external iodide concentra-
tion. Dosimetric calculations revealed only low absorbed doses in
genetically modified tumors (Haberkorn et al. 2001 a, 2003; Sieger
et al. 2003). Therefore, the use of the hNIS gene alone for radioio-
dine therapy of non-thyroid tumors seems questionable, but the
hNIS gene may be used together with 1211, 1241, or even with 99mTc_
---1
500
400
t~-------------
Q)
x 300 I
~
a.
E wild type tumor
hNIS tumor
thyroid
100
o
o 200 400 600 800 1000 1200 1400
Time (min)
Fig. 2. Scintigraphic ally determined iodide uptake (impulses/pixel) in wild-

type hepatomas, hN1S-expressing hepatomas, and the thyroid at different
periods after i.v. administration of 131 1
pertechnetate as a simple reporter system for the visualization of

other genes in bicistronic vectors which allow coexpression of two
different genes.
In comparison to the other reporter genes described, the sodium io-
dide symporter gene may present the advantage that it is not likely to
interact with underlying cell biochemistry and that its substrate per-
technetate is widely available. Iodide is not metabolized in most tis-
sues and, although sodium influx may be a concern, no adverse ef-
fects have been observed to date (Haberkorn et al. 2002). The HSVtk
gene may alter the cellular behavior towards apoptosis by changes in
the deoxynucleotide (dNTP) pool (Oliver et al. 1997), antigens may
cause immunoreactivity, and receptors may result in second messen-
ger activation such as triggering signal transduction pathways. How-
ever, these possible interactions have yet to be studied in detail. For
the D2 receptor system, a mutant gene has been applied which shows
uncoupling of signal transduction (Liang et al. 2001).
122 U. Haberkorn
6.4 Protein-Protein Interaction
Protein interaction analysis delivers information about the possible

biological role of genes with unknown function by connecting them
to other better characterized proteins. Furthermore, it detects novel
interactions between proteins that are known to be involved in a
common biological process and also novel functions of previously
characterized proteins.
Protein interactions have been deduced by purely computational
methods or using large scale approaches (Marcotte et al. 1999;
Tucker et al. 2001). Ideally the characterization of protein interac-
tions should be based on experimentally determined interactions be-
tween proteins that are known to be present at the same time and in
the same compartment. With the increasing availability of intrinsi-
cally fluorescent proteins that can be genetically fused to a wide
variety of proteins, their application as fluorescent biosensors has ex-
tended to dynamic imaging studies of cellular biochemistry even at
the level of organelles or compartments participating in specific pro-
cesses (Wouters et al. 2001). Fluorescence imaging allows the deter-
mination of cell-to-cell variation, the extent of variation in cellular
responses, and the mapping of processes in multicellular tissues. In
addition, procedures for noninvasive dynamic in vivo monitoring are
needed to show whether the protein interactions also work in the
complex environment of a living organism such as mice, rats, or hu-
mans, where external stimuli may affect and trigger cells or organ
function. The yeast two-hybrid technique has been adapted for in
vivo detection of luciferase expression using a cooled charge-
coupled device (CCD) camera (Ray et al. 2002). GAL4 and VPl6
proteins were expressed separately and associated by the interaction
of MyoD and Id, two proteins of the helix-loop-helix family of nu-
clear proteins which are involved in myogenic differentiation. In this
experimental setting, association of GAL4 and VP16 resulted in ex-
pression of firefly luciferase which was under the control of multiple
copies of GAL4 binding sites and a minimal promoter.
Drawbacks of the cooled CCD camera are mainly its limitation to
small animals with different efficiencies of light transmission for dif-
ferent organs, lack of detailed tomographic information, and lack of
an equivalent imaging modality applicable to human studies (Wu et
al. 2001). Therefore, another approach based on the two-hybrid sys-

tem used a fusion of a mutated HSVtk gene and the green fluores-
cent protein gene for in vivo detection of the interaction between
p53 and the large T antigen of simian virus 40 by optical imaging
and PET. Interaction of both proteins resulted in association of
GAL4 and VPl6 and reporter gene expression which was visualized
after administration of an 18F-Iabeled specific substrate for HSVtk
(Luker et al. 2002).
Newer techniques are based on intracistronic complementation and
reconstitution by protein splicing. Complementation does not require
the formation of a mature protein. Both parts of the reporter protein
are active when closely approximated (Rossi et al. 2000). The comple-
mentation strategy can be exploited for a wide range of studies direc-
ted at determining whether proteins derived from two active genes are
coincident or colocalized within cells. Other applications include
transgenic animals expressing complementary lacZ mutants from
two promoters of interest, which should identify cell lines in which
the products of both genes coincide spatially and temporally. Reconsti-
tution is based on protein splicing in trans which requires the reasso-
ciation of an N-terminal and C-terminal fragment of an intein, each
fused to split N- and C-terminal halves of an extein such as a reporter
gene like enhanced green fluorescent protein (EGFP) or luciferase
(Mills and Paulus 200 I). Reassociated intein fragments form a func-
tional protein-splicing active center, which mediates the formation of
a peptide bond between the exteins, coupled to the excision of the
N- and C-inteins. Newer strategies fuse the intein segments to interact-
ing proteins, which results in initiation of protein splicing in trans by
protein-protein interaction (Gimble 1998).
The feasibility of imaging interaction of MyoD and Id based on
both a complementation and a reconstitution strategy has been dem-
onstrated using a CCD camera and split reporter constructs of firefly
luciferase (Paulmurugan et al. 2002). After cotransfection of two
plasmids, the complementation as well as the reconstitution strategy
achieved activities between 40 and 60% of the activity obtained
after transfection of a plasmid bearing the full-length reporter gene.
A cooled CCD camera was applied for visualization of luciferase ac-
tivity in nude mice. This strategy presents a promising tool for the
in vivo evaluation of protein function and intracellular networks and
124 U. Haberkorn
may be extended to approaches involving combinations of reporter

genes and radionuclides. However, MyoD and Id are strong interact-
ing proteins. There may be a weaker in vivo signal when systems
with a weaker interaction are used.
Protein interactions occur not only as physical interactions but
also as functional interactions. These may be studied by analysis of
promoters or promoter modules (Werner 2003) or using combina-
tions of specific promoters and reporter genes (Haberkorn et al.
2002; Haberkorn and Altmann 2003). For the examination of whole
organisms, in vivo reporter systems are promising. These in vivo re-
porters may be used for the characterization of promoter regulation
involved in signal transduction, gene regulation during changes of
the physiological environment, and gene regulation during pharma-
cological intervention. This may be done by combining specific pro-
moter elements with an in vivo reporter gene. However, specific pro-
moters are usually weak (Jiang et al. 2001). This problem was ad-
dressed using a two-step amplification system for optical imaging of
luciferase and PET imaging of HSVtk expression (Iyer et al. 2001).
In that study, tissue-specific reporter gene expression driven by the
prostate-specific antigen promoter was enhanced by the transfer of a
plasmid bearing a GAL4-VP16 fusion protein under the control of
the prostate specific antigen (PSA) promoter together with a second
plasmid bearing multiple GAL4 responsive elements and the repor-
ter gene. Optical imaging revealed a fivefold signal enhancement in
nude mice. Another strategy may be the use of multiple specific en-
hancer elements upstream of their corresponding promoter.
6.S Design of New Biomolecules for Treatment

and Monitoring of New Therapeutic Modalities
Through the accumulation of genomics and proteomics data novel

biomolecules may be discovered or designed. This can be done by
directed genome evolution, metabolic pathway engineering, protein
engineering, analyses of functional genomics and proteomics, high-
throughput screening techniques, and the development of bioprocess
technology (Ryu and Nam 2000). The products of this process will
be monoclonal antibodies, vaccines, enzymes, antibiotics, therapeutic
peptides, and others. The design of a biocatalyst involves two main

steps, which can be iterative: making a set of mutant biocatalysts
and searching the set of mutants with the desired properties. In this
stage of development, isotope-based methods are needed to assess
binding characteristics of antibodies or ligands or measure new
transport or enzyme functions in vivo. Furthermore, in later stages
the coupling of antibodies or peptides with a or f3 emitters may be
used for therapeutic purposes.
Based on the assumption that many therapeutic drugs act through
mechanisms involving perturbations of protein expression, an effi-
cient drug could be defined as one that restores the expression levels
of a ceIl or an organ to the normal state (Anderson and Anderson
1998). Therefore, measurement of the pattern of protein changes can
be used to describe the mechanism of action. Proteomics offers the
opportunity to obtain complimentary information to genomic-based
technologies for the identification and validation of protein targets in
foIlowing time-dependent changes in protein expression levels which
result from selective interaction with specific biological pathways
and identifying protein networks (functional proteomics). Changes in
protein expression or function could also serve as targets for nonin-
vasive imaging procedures. In transgenic mice bearing in vivo repor-
ter genes, protein interactions may be monitored in native organs
and tissues during treatment as wen as during development. Intracel-
lular signaling has been visualized in vitro with combinations of spe-
cific regulatory elements (promoters, enhancers) and reporter genes
such as the secreted alkaline phosphatase (SEAP) downstream of
several copies of specific transcription factor binding sequences
(Ohkubo 2001). A similar approach may be developed for in vivo
detection by use of in vivo reporter genes.
By defining key changes in protein content or function it may be
possible to use radiolabeled ligands or substrates to assess therapy
effects on specific parts of the proteome. This may be done using es-
tablished tracers for new therapies or using new tracers which have
been identified either by functional studies of new genes or by anal-
ysis of changes in expression or functional patterns by the high-
throughput methods of functional genomics. Most studies to date
have been done to assess the effects of suicide gene therapy. Moni-
toring of gene therapy with suicide genes has been done by measure-
126 U. Haberkorn
ments of tumor perfusion and tumor metabolism. Tumor perfusion

in GCV-treated HSVtk-expressing tumors measured after intravenous
administration of [99mTc] hexamethylpropyleneamine oxime
(HMPAO), increased by 206% at day 2 after the onset of GCV treat-
ment (Morin et al. 2000). Also the accumulation of the hypoxia trac-
er [3 H]misonidazole decreased to 34% at day 3, indicating that the
tumor tissue had become less hypoxic during GCV treatment.
The lsF-fluorodeoxyglucose (FDG) uptake has been demonstrated
to be a useful and very sensitive parameter for the evaluation of glu-
cose metabolism during or early after treatment of malignant tu-
mors. Dynamic PET measurements of lsFDG uptake in rats bearing
HSVtk-expressing hepatomas revealed an uncoupling of FDG trans-
port and phosphorylation with enhanced transport values and a nor-
mal phosphorylation rate after 2 days of GCV treatment (Haberkorn
et al. 1998 a). These tumors showed a significant increase of the glu-
cose transporter 1 (GLUT!) as demonstrated by immunohistochem-
istry (Haberkorn et al. 2001 b). The increase in FDG transport nor-
malized after 4 days, whereas the phosphorylation rate increased. As
an underlying mechanism, a redistribution of the glucose transport
protein from intracellular pools to the plasma membrane may be
considered and is observed in cell-culture studies as a general reac-
tion to cellular stress. Consequently, inhibition of glucose transport
by cytochalasin B or competition with deoxyglucose increased apop-
tosis (Haberkorn et al. 2001 b).
Besides these established monitoring procedures, we may expect
new biochemical pathways emerging from proteomics research, lead-
ing to the use of radiolabeled substrates for enzymes, transport sys-
tems, specific structures on cell membranes, or the detection of
apoptosis (Haberkorn et al. 2001 c). For the in vivo detection of
apoptosis, mainly two targets in the apoptotic pathway are of inter-
est, (1) the presentation of phosphatidylserine residues at the outer
side of the plasma membrane and (2) the appearance of activated
caspases (Martin et al. 1995; Villa et al. 1997). Phosphatidylserine is
maintained at the inner site of the plasma membrane by the adeno-
sine triphosphate (ATP)-dependent enzymes floppase and translocase
(Zwaal and Schroit 1997). Apoptosis induced inactivation of these
enzymes and activation of a scramblase leads to the appearance of
phosphatidylserine on the outer side of the membrane. This effect
has been used recently to develop an imaging agent for apoptosis

(Blankenberg et al. 1998, 1999). Annexin V, a 35-kDa human pro-
tein with high affinity for cell membrane bound phosphatidylserine,
was labeled with 99mTc and investigated for its uptake in apoptotic
cells. An increased accumulation was found in lurkat cells where
the programmed cell death was initiated by growth factor depriva-
tion, anti-CD95 antibody, and doxorubicin treatment. Also anti-
CD95-treated mice showed a threefold rise in hepatic 99mTc-Annex-
in V accumulation in response to severe liver damage with histologic
evidence of apoptosis. Finally, increased uptake was detected in ani-
mal models using the acute rejection of transplanted heterotopic car-
diac allografts or transplanted murine B cell lymphomas treated with
cyclophosphamide (Blankenberg et al. 1999).
Since caspases play a key role during the early period of the in-
tracellular signal cascade of cells undergoing apoptosis, benzyloxy-
carbony1-Val-Ala- DL- Asp( 0-methy 1)-fl uoromethy 1 ketone [Z-VAD-
fmk], a pan-caspase inhibitor, was evaluated as a potential apoptosis
imaging agent (Haberkorn et al. 2001 b). Uptake measurements were
performed with Morris hepatoma cells, which showed expression of
the herpes simplex virus thymidine kinase gene. Apoptosis was in-
duced by treatment of the cells with GCV and a twofold increase of
[l3lI]I-Z-VAD-fmk uptake was found at the end of treatment with
the HSVtklsuicide system which constantly remained elevated for
the following 4 hours. The slow cellular influx and lack of uptake
saturation of [131I]lZ-VAD-fmk are evidence for simple diffusion as
transport mechanism. In addition, the absolute cellular uptake of
[131I]IZ-VAD-fmk was found to be low. Instead of using an inhibi-
tor, synthetic caspase substrates are currently being investigated
which may accumulate in the apoptotic cell by metabolic trapping,
thereby enhancing the imaging signal.
6.6 Outlook
Many new molecular structures have been cloned and will be avail-
able as potential novel diagnostic or drug discovery targets. The tar-
get selection and validation will become the most critical component
in this process. The evaluation of genetically manipulated animals or
128 U. Haberkorn
new designed biomolecules will require information about physiol-

ogy, biochemistry, and pharmacology, and the experimental ap-
proaches will apply many technologies including in vivo imaging
with SPECT and PET. Nuclear medicine procedures can be applied
for the determination of gene function and gene regulation using es-
tablished or new tracers, in order to study effects in knockout mice
or in transgenic animals. The measurement of gene regulation may
also be performed using in vivo reporter genes such as enzymes, re-
ceptors, antigens, or transporters. Pharmacogenomics will identify
new surrogate markers for therapy monitoring, which may represent
potential new tracers for imaging. Also, drug distribution studies for
new biomolecules are needed to fasten drug approval in preclinical
stages of drug development. Finally, bioengineering will lead to the
design of new biomolecules by methods such as DNA shuffling or
phage display procedures, which may be used for new approaches in
isotope-based diagnosis and treatment of disease.
References
Alauddin MM, Shahinian A, Kundu RK, Gordon EM, Conti PS (1999) Eval-
uation of 9-[(3- 18F-fluoro-l-hydroxy-2-propoxy)methyl]guanine ([18 F]_
FHPG) in vitro and in vivo as a probe for PET imaging of gene incor-
poration and expression in tumors. Nucl Med Bioi 26:371-376
Anderson L, Seilhamer J (1997) A comparison of selected mRNA and pro-
tein abundances in human liver. Electrophoresis 18:533-537
Anderson NL, Anderson NG (1998) Proteome and proteomics: new technol-
ogies, new concepts, and new words. Electrophoresis 19: 1853-1861
Blankenberg FG, Katsikis PD, Tait IF, Davis RE, Naumovski L, Ohtsuki K,
Kopiwoda S, Abrams Ml, Darkes M, Robbins RC, Maecker HT, Strauss
HW (1998) In vivo detection and imaging of phosphatidylserine expres-
sion during programmed cell death. Proc Natl Acad Sci USA 95:6349-6354
Blankenberg FG, Katsikis PD, Tait JF, Davis RE, Naumovski L, Ohtsuki K,
Kopiwoda S, Abrams MJ, Strauss HW (1999) Imaging of apoptosis (pro-
grammed cell death) with 99mTc annexin V. I Nucl Med 40:184-191
Bogdanov A, Simonova M, Weissleder R (1998) Design of metal-binding
green fluorescent protein variants. Biochim Biophys Acta 1397:56-64
Boland A, Magnon C, Filetti S, Bidart 1M, Schlumberger M, Yeh P, Perri-
caudet M (2002) Transposition of the thyroid iodide uptake and organifi-
cation system in nonthyroid tumor cells by adenoviral vector-mediated
gene transfers. Thyroid 12: 19-26
Cammilleri S, Sangrajrang S, Perdereau B, Brixy F, Calvo F, Bazin H, Mag-

delenat H (1996) Biodistribution of iodine-125 tyramine transforming
growth factor fJ antisense oligonucleotide in athymic mice with a human
mammary tumor xenograft following intratumoral injection. Eur J Nucl
Med 23:448--452
Carlin S, Cunningham SH, Boyd M, McCluskey AG, Mairs RJ (2000) Ex-
perimental targeted radio iodide therapy following transfection of the so-
dium iodide symporter gene: effect on clonogenicity in both two-and
three-dimensional models. Cancer Gene Ther 7:1529-1536
Cho JY, Xing S, Liu X, Buckwalter TLF, Hwa L, Sferra TJ, Chiu 1M,
Jhiang S (2000) Expression and activity of human Na+/I- symporter in
human glioma cells by adenovirus-mediated gene delivery. Gene Ther
7:740-749
Claverie JM (2001) What if there are only 30,000 human genes? Science
291:1255-1257
de Vries EF, van Waarde A, Harmsen MC, Mulder NH, Vaalburg W, Hos-
pers GA (2000) r11C]FMAU and [18F]FHPG as PET tracers for herpes
simplex virus thymidine kinase enzyme activity and human cytomegalo-
virus infections. Nucl Med Bioi 27:113-119
Dewanjee MK, Ghafouripour AK, Kapadvanjwala M, Dewanjee S, Serafini
AN, Lopez DM, Sfakianakis GN (1994) Noninvasive imaging of c-myc
oncogene messenger RNA with indium-Ill-antisense probes in a mam-
mary tumor-bearing mouse model. J Nucl Med 35:1054-1063
Futcher B, Latter GI, Monardo P, McLaughlin CS, Garrels JI (1999) A sam-
pling of the yeast proteome. Mol Cell Bioi 19:7357-7368
Gambhir SS, Barrio JR, Phelps ME, Iyer M, Namavari M, Satyamurthy N,
Wu L, Green LA, Bauer E, MacLaren DC, Nguyen K, Berk AJ, Cherry
SR, Herschman HR (1999) Imaging adenoviral-directed reporter gene ex-
pression in living animals with positron emission tomography. Proc Nat!
Acad Sci USA 96:2333-2338
Gambhir SS, Bauer E, Black ME, Liang Q, Kokoris MS, Barrio JR, Iyer M,
Namavari M, Phelps ME, Herschman HR (2000) A mutant herpes sim-
plex virus type 1 thymidine kinase reporter gene shows improved sensi-
tivity for imaging reporter gene expression with positron emission tomo-
graphy. Proc Nat! Acad Sci USA 97:2785-2790
Gati WP, Misra HK, Knaus EE, Wiebe LI (1984) Structural modifications at
the 2' and 3' positions of some pyrimidine nucleosides as determinants of
their interaction with the mouse erythrocyte nucleoside transporter. Bio-
chern Pharmacol 33:3325-3331
Germann C, Shields AF, Grierson JR, Morr I, Haberkorn U (1998) 5-FJuoro-
J-(2' -deoxy-2' -fluoro-fJ-o-ribofuranosyl)uracil trapping in Morris hepato-
ma cells expressing the herpes simplex virus thymidine kinase gene. J
Nucl Med 39:1418-1423
Gimble FS (1998) Putting protein splicing to work. Chern Bioi 5:R251-
R256
130 U. Haberkorn
Gygi SP, Rist B, Gerber SA, Turecek F, Gelb MH, Aebersold R (1999)
Quantitative analysis of complex protein mixtures using isotope-coded af-
finity tags. Nat Biotechnol 17:994-999
Haberkorn U, Altmann A (2001) Imaging methods in gene therapy of can-
cer. Curr Gene Ther I: 163-182
Haberkorn U, Altmann A (2003) Noninvasive imaging of protein-protein in-
teraction in living organisms. Trends Biotechnol 21:241-243
Haberkorn U, Oberdorfer F, Gebert J, Morr I, Haack K, Weber K, Lindauer
M, van Kaick G, Schackert HK (1996) Monitoring of gene therapy with
cytosine deaminase: in vitro studies using 3H-5-fluorocytosine. J Nucl
Med 37:87-94
Haberkorn U, Altmann A, Morr I, Knopf KW, Germann C, Haeckel R,
Oberdorfer F, van Kaick G (1997 a) Monitoring gene therapy with herpes
simplex virus thymidine kinase in hepatoma cells: uptake of specific sub-
strates. J Nucl Med 38:287-294
Haberkorn U, Altmann A, Morr I, Germann C, Oberdorfer F, van Kaick G
(1997 b) MuItitracer studies during gene therapy of hepatoma cells with
HSV thymidine kinase and ganciclovir. J Nucl Med 38:1048-1054
Haberkorn U, Bellemann ME, Gerlach L, Morr I, Trojan H, Brix G, Alt-
mann A, Doll J, van Kaick G (1998a) Uncoupling of 2-fluoro-2-deoxy-
glucose transport and phosphorylation in rat hepatoma during gene thera-
py with HSV thymidine kinase. Gene Ther 5:880-887
Haberkorn U, Khazaie K, Morr I, Altmann A, Muller M, van Kaick G
(1998 b) Ganciclovir uptake in human mammary carcinoma cells express-
ing herpes simplex virus thymidine kinase. Nucl Med BioI 25:367-373
Haberkorn U, Henze M, Altmann A, Jiang S, Morr I, Mahmut M, Peschke
P, Kubler W, Debus J, Eisenhut M (2001 a) Transfer of the human so-
dium iodide symporter gene enhances iodide uptake in hepatoma cells. J
Nucl Med 42:317-325
Haberkorn U, Altmann A, Kamencic H, Morr I, Traut U, Henze M, Jiang S,
Metz J, Kinscherf R (2001 b) Glucose transport and apoptosis after gene
therapy with HSV thymidine kinase. Eur J Nucl Med 28:1690-1696
Haberkorn U, Kinscherf R, Krammer PH, Mier W, Eisenhut M (2001 c) In-
vestigation of a potential scintigraphic marker of apoptosis: radioiodi-
nated Z-Val-Ala-DL-Asp(O-methyl)-fluoromethyl ketone. Nucl Med BioI
28:793-798
Haberkorn U, Altmann A, Eisenhut M (2002) Functional genomics and pro-
teomics - the role of nuclear medicine. Eur J Nucl Med 29:115-132
Haberkorn U, Kinscherf R, Kissel M, Kubler W, Mahmut M, Sieger S, Ei-
senhut M, Peschke P, Altmann A (2003) Enhanced iodide transport after
transfer of the human sodium iodide symporter gene is associated with
lack of retention and low absorbed dose. Gene Ther 10:774-780
Hamstra DA, Rice DJ, Fahmy S, Ross BD, Rehemtulla A (1999) Enzyme/
prodrug therapy for head and neck cancer using a catalytically superior
cytosine deaminase. Hum Gene Ther 10:1993-2003
Haubner R, Avril N, Hantzopoulos PA, Gansbacher B, Schwaiger M (2000)

In vivo imaging of herpes simplex virus type 1 thymidine kinase gene ex-
pression: early kinetics of radio labelled FrAU. Eur J Nucl Med 27:283-
291
Hospers GA, Calogero A, van Waarde A, Doze P, Vaal burg W, Mulder NH,
de Vries EF (2000) Monitoring of herpes simplex virus thymidine kinase
enzyme activity using positron emission tomography. Cancer Res
60:1488-1491
Hustinx R, Shiue CY, Alavi A, McDonald D, Shiue GG, Zhuang H, Lanuti
M, Lambright E, Karp JS, Eck SL (2001) Imaging in vivo herpes sim-
plex virus thymidine kinase gene transfer to tumour-bearing rodents
using positron emission tomography and (,8 F)FHPG. Eur J Nucl Med
28:5-12
Iversen PL, Zhu S, Meyer A, Zon G (1992) Cellular uptake and subcellular
distribution of phosphorothioate oligonucleotides into cultured cells. Anti-
sense Res Dev 2:211-222
Iyer M, Wu L, Carey M, Wang Y, Smallwood A, Gambhir SS (2001) Two-step
transcriptional amplification as a method for imaging reporter gene expres-
sion using weak promoters. Proc Nat! Acad Sci USA 98: 14595-14600
Jiang S, Altmann A, Grimm D, Kleinschmidt JA, Schilling T, Germann C,
Haberkorn U (2001) Tissue-specific gene expression in medullary thyroid
carcinoma cells employing calcitonin regulatory elements and AAV vec-
tors. Cancer Gene Ther 8:469-472
Kobori N, Imahori Y, Mineura K, Ueda S, Fujii R (1999) Visualization of
mRNA expression in CNS using IIC-Iabeled phosphorothioate oligodeox-
ynucleotide. Neuroreport 10:2971-2974
La Perle KM, Shen D, Buckwalter TL, Williams B, Haynam A, Hinkle G,
Pozderac R, Capen CC, Jhiang SM (2002) In vivo expression and func-
tion of the sodium iodide symporter following gene transfer in the MA-
TLyLu rat model of metastatic prostate cancer. Prostate 50: 170-178
Liang Q, Satyamurthy N, Barrio JR, Toyokuni T, Phelps MP, Gambhir SS,
Herschman HR (2001) Noninvasive and quantitative imaging, in living
animals, of a mutant dopamine D2 receptor reporter gene in which ligand
binding is uncoupled from signal transduction. Gene Ther 19:1490-1498
Loke SL, Stein CA, Zhang XH, Mori K, Nakanishi M, Subasinghe C, Co-
hen JS, Neckers LM (1989) Characterization of oligonucleotide transport
into living cells. Proc Nat! Acad Sci USA 86:3474-3478
Luker GD, Sharma V, Pica CM, Dahlheimer JL, Li W, Ochesky J, Ryan CE,
Piwnica-Worms H, Piwnica-Worms D (2002) Noninvasive imaging of
protein-protein interactions in living animals. Proc Natl Acad Sci USA
99: 6961-6966
MacLaren DC, Gambhir SS, Satyamurthy N, Barrio JR, Sharfstein S, Toyo-
kuni T, Wu L, Berk AJ, Cherry SR, Phelps ME, Herschman HR (1999)
Repetitive non-invasive imaging of the dopamine D2 receptor as a repor-
ter gene in living animals. Gene Ther 6:785-791
132 U. Haberkorn
Mahony WB, Domin BA, McConnell RT, Zimmerman TP (I988) Acyclovir

transport into human erythrocytes. J BioI Chern 263:9285-9291
Mandell RB, Mandell LZ, Link CJ (1999) Radioisotope concentrator gene
therapy using the sodium/iodide symporter gene. Cancer Res 59:661-668
Marcotte EM, Pellegrini M, Ng HL, Rice DW, Yeates TO, Eisenberg D
(1999) Detecting protein function and protein-protein interactions from
genome sequences. Science 285:751-753
Martin SJ, Reutelingsperger CPM, McGahon AJ (1995) Early redistribution
of plasma membrane phosphatidylserine is a general feature of apoptosis
regardless of the initiating stimulus: inhibition by overexpression of BcI-
2 and Abl. J Exp Med 182:1545-1556
Mills KV, Paulus H (2001) Reversible inhibition of protein splicing by Zinc
ion. J BioI Chern 276: 10832-10838
MoncIus M, Luxen A, Van Naemen J, Damhaut P, Velu T, Cool V, Goldman
S (1995) Development of PET radiopharmaceuticals for gene therapy:
synthesis of 9-«I-(18F)f!uoro-3-hydroxy-2-propoxy)methyl)guanine. J La-
bel Compd Radiopharm 37:193-195
Morin KW, Knaus EE, Wiebe LI (1997) Non-invasive scintigraphic monitor-
ing of gene expression in a HSV-l thymidine kinase gene therapy model.
Nucl Med Commun 18:599-605
Morin KW, Knaus EE, Wiebe LI, Xia H, McEwan AJ (2000) Reporter gene
imaging: effects of ganciclovir treatment on nucleoside uptake, hypoxia
and perfusion in a murine gene therapy tumour model that expresses
herpes simplex type-l thymidine kinase. Nucl Med Commun 21:129-137
Mukhopadhyay T, Tainsky M, Cavender AC, Roth JA (1991) Specific inhibi-
tion of K-ras expression and tumorigenicity of lung cancer cells by anti-
sense RNA. Cancer Res 51:1744-1748
Nakamoto Y, Saga T, Misaki T, Kobayashi H, Sato N, Ishimori T, Kosugi S,
Sakahara H, Konishi J (2000) Establishment and characterization of a
breast cancer cell line expressing Na+/I-symporters for radioiodide con-
centrator gene therapy. J N ucl Med 41: 1898-1904
Ohkubo N (2001) Apolipoprotein E4 stimulates cAMP response element-
binding protein transcriptional activity through the extracellular signal-
regulated kinase pathway. J BioI Chern 276:3046-3053
Oliver FJ, Collins MKL, Lopez-Rivas A (1997) Overexpression of a hetero-
logous thymidine kinase delays apoptosis induced by factor deprivation
and inhibitors of deoxynucleotide metabolism. J BioI Chern 272: 10624-
10630
Paulmurugan R, Umezawa Y, Gambhir SS (2002) Noninvasive imaging of
protein-protein interactions in living subjects by using reporter protein
complementation and reconstitution strategies. Proc Nat! Acad Sci USA
99: 15608-15613
Pradet-Balade B, Boulme F, Beug H, Mullner EW, Garcia-Sanz JA (2001)
Translation control: bridging the gap between genomics and proteomics?
Trends Biochem Sci 26:225-229
Raben D, Buchsbaum DJ, Khazaeli MB, Rosenfeld ME, Gillespie GY, Griz-
zle WE, Liu T, Curiel DT (1996) Enhancement of radiolabeled antibody
binding and tumor localization through adenoviral transduction of the hu-
man carcinoembryonic antigen gene. Gene Ther 3:567-580
Ray P, Pimenta H, Paulmurugan R, Berger F, Phelps ME, Iyer M, Gambhir
SS (2002) Noninvasive quantitative imaging of protein-protein interac-
tions in living subjects. Proc Natl Acad Sci USA 99:3105-3110
Rossi FMV, Blakely BT, Blau HM (2000) Interaction blues: protein interac-
tions monitored in live mammalian cells by p-galactosidase complemen-
tation. Trends Cell Bioi 10: 119-122
Ryu DDY, Nam DH (2000) Recent progress in biomolecular engineering.
Biotechnol Prog 16:2-16
Saito Y, Price RW, Rottenberg DA, Fox 11, Su TL, Watanabe KA, Philips
FS (1982) Quantitative autoradiographic mapping of herpes simplex virus
encephalitis with radiolabeled antiviral drug. Science 217:1151-1153
Shi N, Boado RJ, Pardridge WM (2000) Antisense imaging of gene expres-
sion in the brain in vivo. Proc Nat! Acad Sci USA 97:14709-14714
Shimura H, Haraguchi K, Miyazaki A, Endo T, Onaya T (1997) Iodide up-
take and experimental 13 I J therapy in transplanted undifferentiated thy-
roid cancer cells expressing the Na+/I- symporter gene. Endocrinology
138:4493-4496
Sieger S, Jiang S, Schonsiegel F, Eskerski H, Kubler W, Altmann A, Haber-
korn U (2003) Tumour specific activation of the sodium/iodide symporter
gene under control of the glucose transporter gene 1 promoter (GTI-1.3).
Eur J Nucl Med 30:748-756
Simonova M, Weissleder R, Sergeyev N, Vilissova N, Bogdanov A Jr (1999)
Targeting of green fluorescent protein expression to the cell surface. Bio-
chem Biophys Res Commun 62:638-642
Smit JW, Shroder-van der Elst JP, Karperien M, Que I, van der Pluijm G,
Goslings B, Romijn JA, van der Heide D (2000) Reestablishment of in
vitro and in vivo iodide uptake by transfection of the human sodium io-
dide symporter (hNIS) in a hNIS defective human thyroid carcinoma cell
line. Thyroid 10:939-943
Smit JW, Schroder-van der Elst JP, Karperien M, Que I, Stokkel M, van der
Heide D, Romijn JA (2002) Iodide kinetics and experimental 131 1 therapy
in a xenotransplanted human sodium-iodide symporter-transfected human
follicular thyroid carcinoma cell line. J Clin Endocrinol Metab 87: 1247-
1253
Spitzweg C, O'Connor MK, Bergert ER, Tindall DJ, Young CY, Morris JC
(2000) Treatment of prostate cancer by radioiodine therapy after tissue-
specific expression of the sodium iodide symporter. Cancer Res 60:6526-
6530
Spitzweg C, Dietz AB, O'Connor MK, Bergert ER, Tindall DJ, Young CY,
Morris JC (2001) In vivo sodium iodide symporter gene therapy of pros-
tate cancer. Gene Ther 8:1524-1531
134 U. Haberkorn: Future Directions in Molecular Imaging
Stegman LD, Rehemtulla A, Beattie B, Kievit E, Lawrence TS, BJasberg RG,

Tjuvajev JG, Ross BD (1999) Noninvasive quantitation of cytosine deami-
nase trans gene expression in human tumor xenografts with in vivo magnetic
resonance spectroscopy. Proc Natl Acad Sci USA 96:9821-9826
Tjuvajev JG, Stockhammer G, Desai R, Uehara H, Watanabe K, Gansbacher
B, BJasberg RG (1995) Imaging the expression of transfected genes in
vivo. Cancer Res 55:6126-6132
Tjuvajev JG, Avril N, Oku T, Sasajima T, Miyagawa T, Joshi R, Safer M,
Beattie B, DiResta G, Daghighian F, Augensen F, Koutcher J, Zweit J,
Humm J, Larson SM, Finn R, Blasberg R (1998) Imaging herpes virus
thymidine kinase gene transfer and expression by positron emission to-
mography. Cancer Res 58:4333-4341
Tucker CL, Gera JF, Uetz P (2001) Towards an understanding of complex
protein networks. Trends Cell Bioi 11: \02- \06
Urbain JL, Shore SK, Vekemans MC, Cosenza SC, De Riel K, Patel GV,
Charkes ND, Malmud LS, Reddy EP (1995) Scintigraphic imaging of on-
cogenes with antisense probes: does it make sense? Eur J Nucl Med
22:499-504
Villa P, Kaufmann SH, Earnshwa WC (1997) Caspases and caspase inhibi-
tors. TIBS 22:388-393
Watanabe N, Sawai H, Endo K, Shinozuka K, Ozaki H, Tanada S, Murata
H, Sasaki Y (1999) Labeling of phosphorothioate antisense oligonucleo-
tides with yttrium-90. Nucl Med BioI 26:239-243
Werner T (2003) Promoters can contribute to the elucidation of protein func-
tion. Trends Biotechnol 21 :9-13
Wiebe LI, Knaus EE, Morin KW (1999) Radiolabelled pyrimidine nucleo-
sides to monitor the expression of HSV-l thymidine kinase in gene thera-
py. Nucleosides Nucleotides 18: \065-1066
Woolf TM, Melton DA, Jennings CGB (1992) Specificity of antisense oligo-
nucleotides in vivo. Proc Natl Acad Sci USA 89:7305-7309
Wouters FS, Verveer PJ, Bastiaens PIH (2001) Imaging biochemistry inside
cells. Trends Cell BioI 11:203-211
Wu JC, Sundaresan G, Iyer M, Gambhir SS (2001) Noninvasive optical
imaging of firefly luciferase reporter gene expression in skeletal muscles
of living mice. Mol Ther 4:297-306
Zamecnik PC, Stephenson ML (1978) Inhibition of Rous sarcoma virus re-
plication and cell transformation by a specific oligodeoxynucleotide. Proc
Natl Acad Sci USA 75:280-285
Zinn KR, Buchsbaum DJ, Chaudhuri TR, Mountz JM, Grizzle WE, Rogers
BE (2000) Noninvasive monitoring of gene transfer using a reporter re-
ceRtor imaged with a high-affinity peptide radiolabeled with 99mTc or
18 Re. J Nucl Med 41:887-895
Zwaal RF, Schroit AJ (1997) Pathophysiologic implications of membrane
phospholipid asymmetry in blood cells. Blood 89:1121-1132
7 Regulatory Implications -
The EU Perspective
A. Neil
7.1 An Introduction to Regulatory Thinking " . . . . . . . . . . . . , 136

7.2 The Low-Dose Issue. . . . . . . . . . . . . . . . . . . . . . . . . . .. 138
7.3 Regulatory Deliberations on Low-Dose Studies . . . . . . . . . . 141
7.3.1 Scope: Definition of a "Low Dose" . . . . . . . . . . . . . . . . . . 142
7.3.2 General Toxicology . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
7.3.3 Local Tolerance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 144
7.3.4 Genotoxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 144
7.3.5 Additional Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , 145
7.3.6 Compound Classes Covered . . . . . . . . . . . . . . . . . . . . . .. 145
7.4 Comments on the EMEA Microdose Position Paper . . . . . . . 146
7.5 Conclusions................................. 148
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 149
A new guidance document, "Position Paper on the Non-clinical

Safety Studies to Support Clinical Trials, with a Single Microdose"
(CPMP 2003) has been adopted by the Committee for Proprietary
Medicinal Products (CPMP) and released on the home page of the
European Agency for the Evaluation of Medicinal Products
(EMEA). The position document is intended to handle some special
aspects of pre-clinical development, notably prior to imaging studies
in man, where existing guidance has been deemed to be inappropri-
ate. The following chapter is based on the assumption that it would
be of value to researchers - in academia and the pharmaceutical in-
136 A. Neil
dustry alike - to understand the reasoning behind such a document,

as seen by a national regulator within the European Union. Neither
isotope safety, nor marketing applications for imaging agents are,
however, considered here.
7.1 An Introduction to Regulatory Thinking
The regulation of pharmaceuticals within the European Union is not

done through a single organisation, but a collection of national regu-
latory systems held together by - and converging through - common
legal and administrative instruments, namely EU directives that are
implemented in national legislation, as well as common EU guid-
ance documents, incorporating those agreed upon with the Interna-
tional Conference on Harmonisation (ICH) framework. The existing
differences in national regulatory organisations in the EU reflect na-
tional legal and administrative tradition as well as earlier regulatory
experience of where problems are likely to be encountered. This is
very evident in the regulation of clinical trials, where the tradition
varies considerably. The main reason behind the variance is that pa-
tient safety and surveillance of clinical trials, including the fulfil-
ment of internationally agreed requirements for the supporting docu-
mentation that should be available prior to a clinical trial, is solely a
national concern.
As the mandate for the EU regulatory forum, which includes CPMP
and its attached working groups, is weak with respect to the regulation
of clinical trials, the EU regulatory focus in new guidance documents
has been on the data requirements for marketing authorisation (MA)
applications. The documentation required for clinical trials has to a
large degree not being taken into consideration. This state of affairs
will probably change with the new EU Directive 2001120/EC, that im-
plements good clinical practice (GCP) and harmonises the administra-
tive procedure for clinical trial applications across the EU countries
from May 2004 (EU 2001). Decision-making on each application is
left to the member states as before, but the mere fact that different
authorities will be aware of what is going on in other countries, with
a formal right to contact each other, will probably serve as a strong
driving force for future harmonisation.
Regulatory Implications - The EU Perspective 137
All clinical trials are biomedical experiments. As such, they

should be reviewed by an ethics committee (EC) according to inter-
national conventions, such as the European Convention on Human
Rights (Council of Europe 1997) and professional codex as the De-
claration of Helsinki (WMA 2000). The partitioning of institutional
responsibilities in different EU countries is not within the scope of
this chapter, but there will probably always be some overlap in the
assessments made by an EC and a regulatory authority. The human
ethics issues concerning patient dignity and autonomy versus the re-
searchers' and sponsors' interests are clearly a task for the EC. But
it is nevertheless difficult to see how a regulatory authority responsi-
ble for an evaluation of trial product quality, and the scientific basis
for anticipated risks and benefits, could avoid at least a measure of
"technical ethics" in deliberations on the possible benefits versus
risks.
In fact, the benefit/risk deliberation is central to a regulatory ap-
proval of a clinical trial application. It should be realised that each
clinical trial application concerns a prospective, or research, activity
with unknown benefits and risks. This differs from an MA applica-
tion that is based on a substantial amount of retrospective data. A
clinical trial could serve two widely different purposes. It could be
used as a tool for clinical science to investigate pharmacological or
pathophysiological mechanisms that could be used as targets for
pharmaceuticals at a later stage. It could also be used in conven-
tional clinical development of pharmaceuticals to support an MA. Ir-
respective of purpose, a benefit/risk evaluation of a clinical trial ap-
plication must be based on adequate scientific data, and to aid in
this, the requirements for scientifically based non-clinical safety
studies are outlined in a regulatory framework of guidance docu-
ments that have their roots in the OECD guidance on the documen-
tation of chemical toxicity, including requirements for quality assur-
ance through good laboratory practice (GLP).
Concerning risks, the main document is the "ICH topic M3, Tim-
ing of Pre-clinical Studies in Relation to Clinical Trials" (ICH M3
2000) that covers the studies normally needed for appropriate safety
assessments at different stages of clinical development, from prior to
first dose in man until an MA. Although "normally needed" in this
context refers to needs for low molecular weight pharmaceutical
138 A. Neil
compounds, it is important to realise that the ICH M3 document sets

the standard for what properties related to safety should be deter-
mined when during a clinical development program. For other kinds
of products where special guidance documents exist, such as prod-
ucts obtained by biotechnological processes (ICH S6 1997), vaccines
(CPMP 1997), gene therapy products (CPMP 2001), etc., that are fo-
cussed on requirements prior to an MA, this guidance should be
read against the background of the ICH M3.
The benefit term of the assessment is, however, an area where
written non-clinical guidance is weak. It may seem self-evident that
a clinical trial has a purpose, but to what extent the rationale behind
initial studies in man should be backed by pharmacological proof-
of-concept studies in animal models is rarely stated. There are some
exceptions, as the CPMP guidance document on gene therapy
(CPMP 2001) that puts some emphasis on the need for studies of
potential effect.
Guidance, it must be noted, is neither law nor a list of tick-box
items. In principle, any requirement in a guidance document could
be waived if there were proper scientific justification, but it should
be remembered that non-clinical safety studies are also intended as a
safety net to cover for unexpected events. There is also a general
GLP requirement on all safety studies in order to support the accu-
racy, repeatability and accountability of the data.
7.2 The Low-Dose Issue
With the advent of new very sensitive bioanalytical techniques, such

as liquid chromatography coupled to tandem mass spectrometry
(LC-MS/MS), positron emission tomography (PET) imaging, accel-
erator mass spectrometry (AMS), and other very sensitive analytical
techniques, it is now possible to detect picomolar concentrations or
less of individual compounds in tissues or body fluids. It was soon
realised that such powerful tools could be useful in initial drug de-
velopment to characterise a compounds pharmacokinetic/distribution
properties, or receptor selectivity profile, in early clinical trials to se-
lect an appropriate compound or formulation for further develop-
ment. This approach is attractive since it would reduce the number
of clinical trial subjects that are exposed to test material in projects

that will be discontinued at a later stage. Also, the compound expo-
sure in such screens would be low, with doses typically in the low
microgram range or below.
Most early low-dose studies in man were performed by academic
investigators who developed new technologies. If overseen by a reg-
ulatory authority at all, such studies were dealt with on a case-by-
case basis. In view of the low doses used, few trial subjects in-
volved, and the general lack of resources available to the investiga-
tors to perform state-of-the-art toxicology studies, it might be as-
sumed that such trials were accepted based on sub-standard or non-
existent safety data.
So is there a problem? When the number of clinical trials using
the low-dose approach increased and the regulatory authorities
started to evaluate the files, some causes for possible concerns were
noted, especially in relation to PET imaging studies using the la-
belled ligand displacement approach.
- Potent biological effects. Imaging ligands are not any chemical

compounds administered at very low doses. They are developed
to selectively interact with identified pharmacological targets, and
with a high affinity.
- High local concentrations. As compounds used as labelled li-
gands in PET studies usually have a very selective mode of inter-
action, they accumulate to a high degree in certain structures in a
predictive manner. This is the very reason for their use - to ob-
tain the right biological interaction information with a sufficient
signal-to-noise ratio. Thus, although the total dose administered
might be minute, the local concentration could be substantial if
all the material is sequestered, for instance into a few small brain
nuclei.
- History of compounds in the public domain. In the area of ligand
binding studies, the "pharmacologist's toolbox" is full of recep-
tor-type selective reference compounds well known to "the trade"
by their pharmaceutical company codes. The reason why these
compounds were made available for experimental, strictly non-hu-
man, use in the first place may vary, but it may be assumed that
the originator found no commercial use for them. Whether due to
140 A. Neil
lack of intended clinical effect, failure in non-clinical or clinical

safety, or competition in the intended market segment, this infor-
mation is seldom made publicly available. Note that commercial
suppliers of pharmacological reference compounds, labelled or
unlabelled, specify that they are not intended for human use. In
fact, at the Medical Products Agency (MPA), we have come
across a couple of PET ligands that were dropped from commer-
cial development due to failure in genotoxicity or carcinogenicity
assays.
Also, there are well known instances of unexpected CNS adverse ef-
fects occurring from minute doses of compounds, such as the hallu-
cinogenic effect from lysergic acid diethylamide (LSD; see Jaffe
1975), and I-methyl-4-phenyl-l,2,3,6-tetrahydropyridine (MPTP)
neurotoxicity (Bums et al. 1984; Javitch et al. 1984).
Thus, the notion that a very low dose of a compound is always
safe to administer to humans lacks scientific basis. On the other
hand, the risks involved in a single low-dose situation may be con-
sidered slight. The question is then what documentation could form
a minimal basis for a safety evaluation without compromising safety.
Such deliberations need to be put into a wider context.
The primary stakeholder in a clinical trial is always the trial sub-
ject. Not only should there be a sufficient scientific basis for a bene-
fit/risk assessment, where safety is the key factor, but the written in-
formation to trial subjects must also be scientifically correct. There
are secondary considerations as well. Clinical research is extremely
important to society, but it is dependent on trust in the regulatory
system, and any accident occurring that could have been prevented
by adhering to normal requirements could harm clinical research in
general to an unforeseen extent. Also, researchers or industry that
drives the development of new therapies could gain by optimising
the use of available resources. It is therefore important to obtain a
wide consensus on what reductions in ICH M3 requirements on sci-
entific data could be accepted when the perceived risk is minute,
without compromising the safety of the trial subject.
7.3 Regulatory Deliberations on Low-Dose Studies
A regulatory problem was thus identified, namely how to handle the

very restricted number of clinical trial applications where existing
regulatory guidance was clearly not adequate, but in a manner that
neither compromised subject safety nor would lead to ambiguities
and further regulatory problems. Although clinical trials are a na-
tional concern, an EU agreement on common standards is important
to avoid unequal treatment of different competing commercial facil-
ities and to give clear advice to applicants and regulators alike.
Therefore, the issue was raised through the CPMP Safety Working
Party.
As mentioned, the normal requirements for documentation of a
new compound prior to its first use in man is already described in
the ICB M3 document (see Table 1, left column for a summary of
studies involved). Although clearly not intended for, or sometimes
Table 1. Non-clinical documentation required prior to first human dose com-

pared to the qualifications of a microdose study
Normal requirements "Microdose" requirements
Effect (no GLP required) Effect (no GLP)

- Scientific support for effect - Scientific support for study
- Basis for initial dose to man - Dose < 1/100 of active
Safety (rCH M3; GLP) Safety (still GLP)
- Safety pharmacology studies - Safety pharmacology studies
Core battery: CNS; cardiovascular; None (dose < 1/100 of primary effect)
respiration
- General toxicology studies - General toxicology studies
Acute toxicity (2 species) Extended single dose (l species)
Repeat dose toxicity (14 days; i.v. + intended route (limit -dose
2 species) option)
Genotoxicity in vitro - Genotoxicity in vitro
Mutagen and clastogen effects Mutagen and clastogen effects
Reproduction segment II in 2 (Abridged/reduced assays
species accepted)
(Only if fertile women are - Available background data
included)
- Local tolerability
142 A. Neil
even applicable to, the situation of very low doses, the ICH M3 is
an internationally agreed general guideline that cannot be overruled
by new national or EMEA guidance. The issue was therefore how to
come to an agreement within the EU on what deviations from the
ICH M3 recommendations could be regarded as scientifically justi-
fied in the instances when compounds are given as a single very low
dose. This is formally handled as a position paper giving the EU's
view on what data requirements are necessary to support a first, very
low dose in man and could also form the basis for further discussion
within the ICH framework, if needed.
So what are the key issues that need to be resolved to interpret ex-
isting written guidance? What documentation would be indispensable
in order to make a valid safety assessment of a clinical trial application
for a low-dose study? The following points were found crucial.
7.3.1 Scope: Definition of a "Low Dose"
In defining low dose, two interests have to be balanced. Sufficient

doses should be allowed so that researchers can perform meaningful
studies from an analytical point of view, but doses need to be low
enough to reduce the risks to the trial subject through unforeseen ef-
fects. It was decided early on to exclude primary effect studies. To
ensure this, the low dose was to be defined in terms of a fraction of
a pharmacologically effective dose with an upper limit for total
amount of allowable compound. The reason for excluding primary
effects is that compounds potent enough to yield a primary effect at
a very low dose could be suspected to elicit a secondary effect at
the same level of exposure. This point is vital to trial subject safety.
Primary pharmacodynamic effect studies would have to be preceded
by a thorough investigation of compound properties, especially sec-
ondary pharmacology, and it is difficult to justify why such studies
should not be covered by a full non-clinical documentation, accord-
ing to guidance. From a regulatory point of view, a clear distinction
could then be made between studies intended to characterise com-
pound physical-chemical interaction properties, although at very low
doses, and those intended to study biological effects of compounds.
Primary effect screening studies in man are therefore out of the
scope of this CPMP position paper, since it was not aimed at reduc-
ing the non-clinical data requirements, as outlined in ICH M3 for
such studies. To this author, it is not even clear if such an approach
for effect studies would fulfil the requirements of "adequately per-
formed laboratory and animal experimentation" and "careful assess-
ment of predictable risks in comparison with foreseeable benefits",
as expressed in the Declaration of Helsinki (WMA 2000). Moreover,
adding primary effect studies would involve such a change from ex-
isting guidance that it should be handled as a new or major ICH top-
ic revision before being put into use.
7.3.2 General Toxicology
General toxicology is the area where the discrepancy between the

regulatory requirements as stated in the ICH M3 and the intended
clinical use supported by the studies is most evident and, also,
where regional regulatory requirements differ. In the United States,
Monro and Mehta (1996) raised the issue, proposing that single-dose
studies, focussing on secondary pharmacology rather than histo-
pathological changes, would better support early single-dose studies
in man. The FDA quickly responded to this proposal (Choudary et
al. 1996) and introduced its guidance for industry on "Single Dose
Acute Toxicity Testing for Pharmaceuticals" (FDA 1996). This guid-
ance allows for early explorative studies, not only at low doses. Cor-
responding guidance is lacking in the EU, and an approach between
the United States and the EU practices would be attractive for har-
monisation purposes. A proposal for a single-dose toxicity study as
a sufficient basis for early pharmacokinetic investigations has also
been made from individual EU regulators (Spindler et al. 2000).
A change from 2 weeks repeat-dose toxicity in two species to an
extended single-dose study by the intravenous route in one well-jus-
tified species would yield a substantial economy in terms of animals
and test compound. However, the U.S. single-dose protocol still re-
quires the dosing of animals up to a maximum tolerated dose
(MTD), or grams per kilogram body weight if no MTD is reached.
As a basis for a microgram study in man, this might seem excessive.
Clearly an upper "limit dose", with a sufficient dose margin to a no
144 A. Neil
observed adverse effect level (NOAEL) to ensure trial subject safety,

could be used.
To agree on the size of limit-dose margins and their definition is,
however, not trivial. It should be remembered that, in the normal
case, pharmaco-toxicological assessments are based on the identifi-
cation of toxicological properties, and the setting of acceptable
safety margins in relation to (1) perceived risks in terms of their re-
levance to man, (2) consequences, (3) likely frequency, (4) early
signs of danger to the individual, and (5) opportunities to inter-
vene ... In other areas of applied toxicology, safety factors, often as
multiples of 10, are used to cover for various uncertainties. Relevant
factors identified in the current situation are species differences in
sensitivity, unknown severity of toxic effects, and the absence of an
observed effect at a limit dose based on results from a few indivi-
duals. To this, allometric scaling of doses based on body surface
rather than weight should be added, as used in other regulatory
documents for prediction from animals to man (see ICH Q3C 1997
for factors).
7.3.3 Local Tolerance
Local tolerance would best be covered by a single-dose tOXICIty

study that uses the intended route of administration in those cases
where the - default - intravenous route is not sufficient.
7.3.4 Genotoxicity
Most new compounds contemplated for clinical trials stem from in-
dustrial research projects where genotoxicity screens, when appropri-
ate, are applied at an early stage to filter out possible problem com-
pounds. To screen on a large scale, new methods have been devel-
oped, both by commercial vendors and within the pharmaceutical in-
dustry, using very little compound, but the methods are sufficiently
validated against the conventional regulatory assays to be considered
reliable. Data from such assays should be acceptable as a basis for
the safety assessment of compounds belonging to well-known
classes where prior genotoxicity data exist and no structure alerts are
present. Note that if screening assays are used, validation data
should be provided, and the assays should be sufficiently documen-
ted as to be reconstructible, i.e., GLP-compliant.
7.3.5 Additional Data
Except for the primary pharmacodynamics data involved, data from

studies that are performed during the drug discovery phase are rarely
submitted to regulatory authorities. The main reason is that they
usually are superseded by formal pre-clinical safety assessment stud-
ies performed according to GLP. In a situation when a regulator is
to assess an application for a clinical trial on a reduced set of data,
it would be of value if the regulator could obtain an insight into the
information used as a basis for the decision to select one compound
over another for the trial. Such additional information could never
be a sole basis for an authoritative evaluation, but could clarify po-
tential issues before they are raised. Examples are safety screens
used and (de-)selection criteria applied in these, as well as any data
obtained for close analogues in safety studies.
7.3.6 Compound Classes Covered
The discussion on low-dose studies in man has been focussed on

low molecular weight organic compounds. These are the ones most
frequently used for labelling in imaging studies, and their distribu-
tion and metabolic properties are most variable and, therefore, in
need of early pharmacokinetic investigations. Biotechnologically de-
rived compounds have special properties that are thought best
handled on a case-by-case basis rather than within a general com-
ment added to an existing guideline. They are, generally, much more
species-specific and, often, have very potent biological effects. More
information than for low molecular weight compounds may be re-
quired to determine if "low-dose" requirements are fulfilled. Also, if
any immunological risks are involved, even with very low parenteral
doses, they may have to be assessed on an individual basis as de-
146 A. Neil
scribed in ICH S6 (1997). For cytotoxic alkylating anticancer com-

pounds there are special safety concerns that may require a separate
approach (CPMP 1998).
7.4 Comments on the EMEA Microdose Position Paper
Agreeing on a common minimal standard takes time. The issue of a

common EU approach to handle low doses in clinical trials was
raised in the CPMP Safety Working Party in June 2001. A concept
paper stating the scope of a future position paper was published by
the CPMP in February 2002, a draft position paper was released for
comments in June 2002, and the final document was adopted by the
EMEA in January 2003, coming into operation by July 2003. A
summary of the changes in comparison to the ICH M3 guideline is
given in Table 1.
There are several features worth noting in the final document. To
start, the title was re-phrased as "Position Paper on Non-clinical
Safety Studies to Support Clinical Trials with a Single Microdose",
emphasising that repeat-dose studies are not within its scope, and
that only studies with very low amounts of active compound are cov-
ered. Further, the term "microdose" is defined as less than III OOth of
the dose calculated to yield a pharmacological effect "based on pri-
mary pharmacodynamic data". To further delimit the "microdose"
concept, and minimise the risk for unforeseen effects, a maximum
human dose of 100 Ilg was specified. It must be emphasised that a
100-llg dose is not a default "cut-off' level below which no data for
compounds are required. Pharmacodynamic activity must be docu-
mented for all compounds. The reason to exclude pharmacological
effect was discussed earlier. A factor of 100 intended to achieve this
objective may seem arbitrary and lacking a scientific basis. But it
should, instead, be regarded primarily as a demand for thorough ap-
praisal of the potentially effective dose in man, with some margin
for species variability. In theoretical terms, for any ligand-protein in-
teraction obeying the law of mass action, a two order of magnitude
concentration difference would cover most of the log dose-binding
curve (Goldstein et al. 1974). For a receptor antagonist, this would
translate to a decrease from about 90% to about 10% receptor occu-
pancy when the ligand concentration is decreased to 111 ~Oth. For

agonists, the dose-response curves are generally steeper, due to am-
plification mechanisms (see Stephenson 1956 for the classical exam-
ple). There is thus some theoretical backing for the notion that a
100-fold difference in administered dose should be sufficient to
avoid a pharmacodynamic effect - provided inter-species variability
is checked for by comparative in vitro studies. In the position paper,
there is also a comment on the development stage, where the under-
lying assumption is that the studies should be "pre-phase I". This
phrasing is ambiguous, and a pre-phase I is not defined in regulatory
terms, but it might be assumed that single-dose studies where a
"microdose" imaging compound is used on a routine basis in a num-
ber of projects over the years is not foreseen. An in vivo diagnostic
used on a routine basis would need an MA, irrespective of doses or
detection technologies used. In fact, at least one SPECT ligand is
currently on the market.
The really new approach is in the general toxicology study re-
quirements, where a single-dose toxicity study in one species could
be accepted to replace the ICH M3 recommendations for safety
pharmacology, as well as acute and repeat toxicity studies in two
species. However, the choice of this species must be justified by
comparative ADME and pharmacodynamic data obtained in vitro.
The focus here is, again, on obtaining reasonably relevant informa-
tion instead of "tick-box" studies. Hence, the need for "sufficiently
large" dose groups is also emphasised.
To this single-dose study, an option for a limit-dose approach has
been added for reasons discussed earlier. The agreed safety factor to
be applied is 1,000. There is one order of magnitude allowed each
for: the variability in sensitivity between species, unknown toxicity,
and small group data. On top of this, the appropriate allometric scal-
ing from animal species to man should be added, 5 for rat, 12 for
mouse, etc., according to (ICH Q3C 1997). For example, if the in-
tended total human dose is 5 ~g, or 0.1 ~g/kg, and, taking the limit-
dose approach, rats are given 5, then 5 (for rat) x 1,000xO.l ~g/
kg=O.5 mg/kg, and barring any adverse findings, the study would be
acceptable.
148 A. Neil
7.5 Conclusions
Clinical trials do differ from marketing authorisations in that clinical

trials are biomedical research activities involving unknown benefits
and risks where the prime concern is safety to the individual subject,
which is strictly a national issue in the EU. There is a common set
of guidelines on different aspects of clinical trials, but the interpreta-
tion resides solely with the different member states. It is likely that
clinical trials will remain a national concern in the EU for the fore-
seeable future, but also that the regulatory handling will converge
with the implementation of the clinical trials directive (EU 2001) in
May 2004. A first step towards a common interpretation of existing
guidelines has been taken to fill a well-defined regulatory problem
concerning very low, microgram doses given to human subjects to
study pharmacokinetic properties or molecular interactions of a com-
pound not intended to cause a pharmacological response. The driv-
ing force for this has been the obvious lack of proportion between
ICH guidance requirements and the perceived risks involved for the
trial subject in these types of trials.
To extend the "microdose" approach to allow for early pharmaco-
dynamic exploratory or screening studies based on similarly reduced
non-clinical data is not a simple matter since the existing ICH M3
guidance is written also to cover pharmacodynamic-effect studies
(see staging definitions in ICH E8 1997). Such an extension of the
term microdose, or any low-dose concept, to cover pharmacody-
namic studies would amount to more than interpretation - new or al-
tered regulatory guidance may have to be agreed upon within the
ICH framework. However, replacing a 14-day repeat-dose toxicity
study with an extended single-dose study that follows the FDA re-
commendations may be found acceptable to support single-dose
pharmacodynamic screening studies in man, provided that such data
are backed by solid safety pharmacology studies.
Biotechnologically derived products and cytotoxic anti-cancer
compound have special properties that may best be handled on a
case-by-case basis rather than within the framework of comments to
existing guidelines.
References
Bums RS, Markey SP, Phillips JM, Chiueh CC (1984) The neurotoxicity of
I-methyl-4-phenyl-l,2,3,6-tetrahydropyridine in the monkey and man.
Can J Neurol Sci 11(1 Suppl):166-168
Choudary J, Contrera JF, DeFelice A, DeGeorge 11, Farrelly JG, Fitzgerald
G, Goheer MA, Jacobs A, Jordan A, Meyers L, Osterberg R, Resnick C,
Sun CJ, Temple RJ (1996) Response to Monro and Mehta proposal for
use of single-dose toxicology studies to support single-dose studies of
new drugs in humans. Clin Pharmacol Ther 59:265-267
Council of Europe (1997) Convention for the protection of human rights and
dignity of the human being with regard to the application of biology and
medicine: Convention on Human Rights and Biomedicine, ETS No 164.
http://conventions.coe.intitreatyIEN/cadreprincipal.htm
CPMP (1997) Note for guidance on preclinical pharmacological and toxico-
logical testing of vaccines (CPMP/SWP/565/95 adopted December 97).
http://www.emea.eu.intipdfs/humaniswp/046595en.pdf
CPMP (1998) Note for guidance on the pre-clinical evaluation of anticancer
medicinal products (CPMP/SWP/997/96 adopted July 1998). http://
www.emea.eu.int/pdfs/human/swp/099796en.pdf
CPMP (2001) Note for guidance on the quality, preclinical and clinical as-
pects of gene transfer medicinal products (CPMP/BWP/3088/99 adopted
April 2001). http://www.emea.eu.int/pdfs/humanlbwp/308899en.pdf
CPMP (2003) Position paper on the non-clinical safety studies to support
clinical trials, with a micro dose (CPMP/SWP/2599/02 adopted January
2003). http://www.emea.eu.int/pdfs/human/swp/259902en.pdf
EU (2001) Directive 2001l20lEC of the European Parliament and of the
Council of 4 April 2001 on the approximation of the laws, regulations
and administrative provisions of the Member States relating to the imple-
mentation of good clinical practice in the conduct of clinical trials on
medicinal products for human use. Official Journal 1.5.2001, L 121/34.
http://pharmacos.eudra.orgIF2/eudralexivol-lInew_v l/Dir200 1-20_en. pdf
FDA (1996) United States Food and Drug Administration guidance for in-
dustry on single dose acute toxicity testing for pharmaceuticals (August
1996). http://www.fda.gov/cder/guidance/ptl.pdf
Goldstein A, Aronow L, Kalman SM (1974) Molecular mechanisms of drug
action: consequences of drug receptor interaction. In: Principles of drug
action: the basis of pharmacology, 2nd edn. Wiley, New York, pp 82-111
ICH E8 (1997) ICH topic E8, general considerations for clinical trials
(CPMP/ICH/291195 adopted September 1997). http://www.emea.eu.inti
pdfs/humaniich/029196en.pdf
ICH M3 (2000) ICH topic M3, timing of pre-clinical studies in relation to
clinical trials: (CPMP/ICH1286/95 Mod. released November 2000). http://
www.emea.eu.int/pdfs/humaniich/028695en.pdf
150 A. Neil: Regulatory Implications - The EU Perspective
ICH Q3C (1997) ICH topic Q3C, note for guidance on impurities: residual
solvents (CPMPIICHl283/95 adopted September 1997). http://
www.emea.eu.int!pdfs/human/ich/028395en.pdf
ICH S6 (1997) ICH topic S6, safety studies for biotechnological products
(CPMPIICHl302/95 - adopted September 97). http://www.emea.eu.int!
pdfs/humaniich/030295en.pdf
Jaffe JH (1975) Drug addiction and drug abuse. In: Goodman LS, Gilman A
(eds) The pharmacological basis of therapeutics, 5th edn. Macmillan,
New York, pp 309-310
Javitch JA, Uhl GR, Snyder SH (1984) Parkinsonism-inducing neurotoxin,
N-methyl-4-phenyl-l,2,3,6-tetrahydropyridine: characterization and local-
ization of receptor binding sites in rat and human brain. Proc Nat! Acad
Sci USA 81:4591-4595
Monro A, Mehta D (1996) Are single-dose toxicology studies in animals
adequate to support single doses of new drug in humans? Clin Pharmacol
Ther 59:258-264
Spindler P, Sjoberg P, Knudsen LE (2000) First exposure in man: toxicologi-
cal considerations. Pharmacol Toxicol 86(Suppl 1):8-12
Stephenson RP (1956) A modification of receptor theory. Br J Pharmacol
11:379-393
WMA (2000) World Medical Association, Declaration of Helsinki 2000.
http://www.wma.net!
8 Nonclinical Development
of Radiopharmaceuticals: Regulatory
Considerations for the United States Food
and Drug Administration
S. Wilson
8.1 Introduction................................. 152

8.2 U.S. Regulatory Considerations for the Nonclinical
Development of Radiodiagnostics . . . . . . . . . . . . . . . . . . . 152
8.2.1 Pharmacology and Safety Pharmacology Studies. . . . . . . . .. 153
8.2.2 Pharmacokinetic and Biodistribution Studies; Dosimetry. . . .. 154
8.2.3 Toxicology Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 155
8.2.4 Genotoxicity Studies. . . . . . . . . . . . . . . . . . . . . . . . . . .. 156
8.2.5 Reproductive Toxicity Studies. . . . . . . . . . . . . . . . . . . . .. 157
8.2.6 Special Toxicology Studies . . . . . . . . . . . . . . . . . . . . . . . 157
8.2.7 Dose and Formulation Considerations for Nonclinical Studies. 158
8.2.8 Development of Previously Approved or Previously Tested
Drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
8.3 U.S. Regulatory Considerations for the Nonclinical
Development of Radiotherapeutics . . . . . . . . . . . . . . . . . .. 160
8.3.1 Evaluation of Potential Radiation Effects
of a Radiotherapeutic . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
8.3.2 Considerations for Conduct of a Toxicity Study
with Drug/Radionuclide . . . . . . . . . . . . . . . . . . . . . . . . . 163
8.4 Conclusion.................................. 163
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 164
152 S. Wilson
8.1 Introduction
In general, the nonclinical studies conducted to support the develop-

ment of radiopharmaceuticals should be consistent with current Inter-
national Conference on Harmonization (ICH) Guidelines. The United
States Food and Drug Administration (FDA), however, has recognized
that certain characteristics of radiopharmaceuticals necessitate an
adaptation of the type and scope of the non clinical studies that are typ-
ically utilized in the development of drugs. From a regulatory perspec-
tive, the nonclinical studies that are recommended by the FDA to sup-
port the development of radiopharmaceutical drugs are dependent on
the intended indication of the drug as either a radiodiagnostic or a
radiotherapeutic. There is an overlap of safety and toxicology issues
for radiodiagnostics and radiotherapeutics because of the presence of
the radionuclide. There are, however, usually significant differences
between the two general categories of radiopharmaceuticals with re-
spect to dose, frequency and duration of dosing, and pharmacological
activity. Consequently, the regulatory considerations for the nonclini-
cal development of radiodiagnostics are considerably different from
radiotherapeutics. The FDA has recently released a revised draft guid-
ance document ("Guidance for Industry: Medical Imaging Drug and
Biological Products - Part I: Conducting Safety Assessment - May
2003") that specifically addresses the development of medical imag-
ing agents including radiodiagnostics. At this time, however, there is
no FDA or ICH guidance that expressly outlines considerations for
the nonclinical studies to support the development of radiotherapeu-
tics. The following discussion will address regulatory considerations
for nonclinical studies for radiodiagnostics and radiotherapeutics sep-
arately, describing the recommended studies, timing of the studies, and
the rationale for the recommended studies.
8.2 U.S. Regulatory Considerations

for the Nonclinical Development of Radiodiagnostics
The studies that the FDA requires to support initiation of clinical

trials with respect to the types and timing of the nonclinical studies
are consistent with ICH Guideline M3: "Nonclinical Safety Studies
Nonclinical Development of Radiopharmaceuticals 153
for the Conduct of Human Clinical Trials for Pharmaceuticals" (ICH

1997 a). There are several considerations for diagnostic radiopharma-
ceuticals, apart from safety issues stemming from the presence of
the radioactive moiety, which will impact the scope and design of
the studies. The considerations include a general lack of pharmaco-
logical activity, the low dose or mass typically administered in the
clinic, proposed clinical use (i.e., single dose), and the biological,
physical, and effect half-lives, which tend to be relatively short. The
studies outlined in Sects. 8.2.1-8.2.6 essentially represent a core bat-
tery of studies for a nonbiological radiodiagnostic. Additional stud-
ies may be required on a case-by-case basis. Nonclinical evaluation
of biological radiodiagnostics should be consistent with ICH Guid-
ance S6: "Preclinical Safety Evaluation of Biotechnology-Derived
Pharmaceuticals" (ICH 1997 c) and will not be described.
8.2.1 Pharmacology and Safety Pharmacology Studies
There are no specific regulatory requirements for pharmacology stud-

ies because radiodiagnostics typically have little or no pharmacologi-
cal activity. Generally, however, pharmaceutical companies will con-
duct studies prior to initiation of phase I clinical trials to provide
proof of concept and a rationale for assuming that the radiodiagnostic
will be an effective medical imaging agent. With the advent of molec-
ular targeting, however, there is a potential for pharmacological activ-
ity to be associated with administration of the radiodiagnostic. If, for
example, the drug binds to a receptor or enzyme, the FDA will want
to see data that define the potential for the pharmacologic and physi-
ologic effects associated with the interaction of the drug with its target.
When a drug is pharmacologically active, the total mass dose, i.e., the
unlabeled and labeled portions, are important and the amount of unla-
beled drug and the relative contribution to the pharmacological activity
should be identified. It is important to keep in mind that a drug may
exhibit species specificity with respect to binding. The drug, there-
fore, should be pharmacologically active in the species used in the con-
duct of the pharmacology studies.
The standard battery of safety pharmacology studies, as outlined
in ICH Guideline S7 A: "Safety Pharmacology Studies for Human
154 s. Wilson
Pharmaceuticals" (ICH 200 I) should be conducted prior to phase I

for a radiodiagnostic. In addition to the core studies to assess effects
of the drug on the cardiovascular (CV), respiratory, and central ner-
vous (CNS) systems, studies should also be conducted to identify
potential perturbations in any organ system in which the drug is in-
tended to accumulate, i.e., organs to be imaged. Even if the studies
are not conducted prior to phase I, it is likely that the studies will be
requested by the FDA prior to approval, if not prior to initiation of
phase II or III studies, to ensure that the nonclinical database is
complete. Even though the CV and respiratory systems and CNS can
be assessed in the clinical setting, evaluation of the radiodiagnostic
in healthy animals and at considerably higher doses than will be
used in humans provides a more rigorous evaluation of the drug.
8.2.2 Pharmacokinetic and Biodistribution Studies; Dosimetry
A critical part of the safety evaluation of a radiopharmaceutical that

must be conducted prior to initiation of any clinical trials is an as-
sessment of radiation exposure or internal radiation dosimetry. Dosi-
metry, a method that calculates the amount and distribution of the
radionuclide, is an integral part of calculating the potential risk of
administration of a radiopharmaceutical. Pharmacokinetic studies, in-
cluding the biodistribution studies, should be conducted in at least a
single species with radiolabeled or "hot" drug prior to phase I. The
study should generate adequate data to calculate the whole-body ab-
sorbed radiation dose as well as the radiation dose to the major or-
gans and tissues, targeted organs and tissues, and adjacent organs of
interest. Ideally, biodistribution studies should be conducted in both
species in which the toxicity studies are conducted so that the distri-
bution data can be correlated with any identified toxicity. The study
design for the nonclinical study should take into consideration the
intended popUlation. For example, if the data are to be used to sup-
port a pediatric indication, all organs rather than selected organs
should be evaluated.
Administration of radiodiagnostics is typically by the intravenous
route and, consequently, concerns with adequate exposure are less of
an issue than for drugs administered by other routes such as oral and
subcutaneous. Regardless, standard pharmacokinetics (PK) or toxico-

kinetics (TK) parameter should be calculated, with the analysis
based on samples from the toxicity studies. Additional PK and meta-
bolic data should be generated prior to initiation of phase II trials, as
is the case for other nonbiologic drugs. Although not required by the
FDA, the metabolism data should be generated early in the drug de-
velopment process to ensure that the appropriate species are used in
the nonclinical toxicity studies. Characterization of the metabolic
profile allows a qualitative and quantitative comparison between the
animal species and humans.
8.2.3 Toxicology Studies
Many of the principles of study design for nonclinical toxicity studies

intended to support the clinical use of radiodiagnostic drugs are similar
to other drugs. For example, the route of administration, treatment reg-
imen, and duration of dosing should be the same as that intended for
the clinical trial, and the studies should be conducted in both a rodent
and nonrodent species. Because radiodiagnostics are typically admi-
nistered as a single dose, the FDA will accept a single-dose toxicity
study to support a single-dose phase VII clinical trial. In fact, the
FDA encourages conduct of a single-dose nonc1inical study in both
a rodent and nonrodent species because the design more closely ap-
proximates the proposed clinical usage. One of the reason the FDA
encourages conducting an acute dose study to support a first in human
study is that the starting dose may be higher when the dose is based on
the no observed adverse effect level (NOAEL) from a single-dose
study instead of a 14- to 28-day repeat-dose toxicity study.
The recommended study design for an expanded acute study is
briefly outlined in the FDA guidance: "Single Dose Acute Testing
for Pharmaceuticals" (FDA 1996). As for a standard acute study, the
total in-life phase is 14 days. However, there is an interim sacrifice
period at approximately 3-4 days after dosing in addition to the ter-
minal sacrifice. The endpoints are also more comprehensive in the
expanded acute study compared to the acute study, and are compar-
able to the endpoints typically included in repeat dose toxicity, e.g.,
clinical pathology, organ weights, and histopathology.
156 s. Wilson
While the duration of the in-life phase for an expanded acute is

comparable to a standard acute study, the time required to generate
a report, the cost, and the number of animals required approaches or
exceeds that of a 14- to 28-day repeat-dose study. The amount of
test article required to conduct the expanded acute study, however, is
significantly reduced compared to a repeat dose study. Availability
of test article is often a major consideration early in the develop-
ment of a drug and may be a rate limiting consideration in develop-
ment. It should be noted that even if a 14- to 28-day repeat-toxicity
study and a standard acute toxicity study were used to support the
initial clinical trial, it is likely that the FDA will request an ex-
panded acute study in a rodent and nonrodent species prior to initia-
tion of phase II clinical trials. If only an expanded acute toxicity
study is conducted prior to phase I, a standard 14- to 28-day repeat-
dose study in a rodent and nonrodent species may be conducted
prior to phase II clinical trials. The FDA will expect to see a 28-day
repeat-dose toxicity study in two species prior to New Drug Applica-
tion (NDA) submission.
8.2.4 Genotoxicity Studies
While it is recognized that the presence of the radionuclide in a

medical imaging agent represents a genotoxic risk, the FDA recom-
mends that the standard battery of genotoxicity studies be conducted
in accordance with the ICH guidance S2B: "Genotoxicity: A Stan-
dard Battery for Genotoxicity Testing of Pharmaceuticals" (ICH
1997 b). Specifically, the in vitro assays to assess the mutational and
clastogenic potential of the drug (Ames assay and mammalian chro-
mosomal aberration assay, respectively) should be conducted prior to
phase I, while the in vivo mouse or rat micronucleus should be con-
ducted prior to phase II. The rationale is that, separate from the ra-
diation effect, the drug may be potentially genotoxic, and this poten-
tial could be additive to or synergistic with the genotoxic effects of
the radionuclide. The FDA guidance for developing imaging agents
indicates that a request for waiver of the studies, including the scien-
tific justification for the request, would be considered. It is likely,
however, that as a minimum the in vitro assays will be required.
Conduct of the in vitro studies represents a very small segment of

the overall development program with respect to time and cost con-
siderations. Because the radionuclide is considered a genotoxin, it is
unlikely that a positive result in the in vitro assays would signifi-
cantly impact the development of the radiodiagnostic.
8.2.5 Reproductive Toxicity Studies
The FDA has generally waived the requirement for fertility and
early embryonic development (segment I) and pre and postnatal
(segment III) reproductive toxicity studies. The rationale was based
on the intended clinical use and characteristics of a radiodiagnostic,
i.e., single dose, low dose, and generally no pharmacological activ-
ity. Until recently, however, the agency has required conduct of tera-
togenicity (segment II) studies in two species, the rat and the rabbit.
The position outlined in the current draft guidance for development
of imaging agents states that the studies can be waived based on
"adequate scientific justification", consistent with the Code of Feder-
al Regulations (2002; 21 CFR 312.10 and 314.90). If a formal re-
quest for waiver is granted by the FDA, the lack of animal reproduc-
tive toxicity data will be noted in the label. In addition, the label
will typically include statements that the potential harm to the fetus
or reproductive capacity is unknown and that the radiodiagnostic
should only be used if the benefit outweighs the risk.
8.2.6 Special Toxicology Studies
Because radiodiagnostics are typically administered by the intrave-

nous route, special toxicology studies are usually recommended
prior to phase I and typically are designed to address concerns per-
taining to local tolerance and protein and red blood cell compatibil-
ity. Local tolerance studies may be addressed in the standard toxicol-
ogy studies if the injection site is evaluated histopathologically. It is
not unusual, however, for dedicated local tolerance studies to be
conducted in which effects at the injection site following intentional
extravasation are evaluated. Whole blood compatibility can be evalu-
158 s. Wilson
ated in vitro using protein flocculation and hemolysis assays. The
need for the studies is in part dependent on the volume of drug that
will be administered in the clinical setting. Whole blood compatibil-
ity is generally more of a concern for nonradioactive medical imag-
ing agents, which are administered at larger volumes than radiodiag-
nostics.
8.2.7 Dose and Formnlation Considerations

for Nonclinical Studies
Until relatively recently, the FDA has recommended that the high
dose used in the nonclinical toxicology studies either elicit signifi-
cant toxicity or represent the maximum feasible dose (MFD). Conse-
quently, because the clinical dose of radiodiagnostic drugs is gener-
ally very low and because radiodiagnostic drugs generally are not
pharmacologically active, it was not unusual to see nonclinical stud-
ies in which the high dose was several 1,000-fold times higher than
the human exposure, based on both a mass and surface area basis.
The high dose was often the MFD rather than a dose that demon-
strated significant toxicity. A similar scenario occurred with the
safety pharmacology studies with respect to multiples of the antici-
pated maximum human exposure. While the draft guidance for de-
velopment of medical imaging agents does not provide a statement
that specifically provides criteria for setting the doses in the noncli-
nical studies, the discussion of the safety-margin criteria for deter-
mining whether a radiopharmaceutical is a Group 1 or 2 drug pro-
vides insight into what the FDA considers reasonable multiples of
the human exposure.
Radiodiagnostics that are identified as Group 1 drugs are consid-
ered to have a very low potential for toxicity. Group 1 designation
will generally not significantly alter the recommended nonclinical
studies, but the scope of the clinical trials necessary to support an
NDA will tend to be less comprehensive with respect to safety moni-
toring than for a Group 2 drug. The safety margin criteria for a
Group 1 designation is that the NOAEL in the acute and safety phar-
macology studies is at least 100 times greater than the proposed
maximal clinical dose (either on a surface area or AUC basis) or
that the NOAEL in a 28-day repeat-dose toxicity study is at least 25

times greater than the maximal clinical dose. A high dose that is
considered to be the NOAEL and is 100 to 200 times the proposed
clinical dose based should, therefore, be adequate. The dose that is
the anticipated maximum recommended for humans, however, may
change, i.e., increase, through the course of drug development.
Therefore, it is advisable to factor the possibility that the recom-
mended human dose may increase when setting the high dose for
the pivotal toxicity studies.
The test article used in the nonclinical toxicity studies, e.g., safety
pharmacology, toxicology, and genetic toxicology studies, should be
degraded or "cold" drug. Theoretically, use of degraded drug could
introduce degradants at levels that are not clinically relevant and
could impact the outcome of the toxicity study. While theoretically
possible, in practice the concern has not been a significant issue. Ide-
ally, the formulation used in the nonclinical studies should also be
comparable or identical to that used in the clinical trials. It may, how-
ever, be necessary to use a more concentrated formulation in the non-
clinical studies to achieve adequate multiples of human exposure.
It is recognized that formulations frequently are modified as drug
development proceeds. If the change is not substantive, additional
nonclinical toxicology studies may not be needed. If, however, an
excipient is used that has not been previously tested, used in an ap-
proved product, administered by the proposed route, or administered
at the same dose level, or if the formulation change is anticipated to
impact the biodistribution of the drug, qualification of the excipient
and bridging toxicity along with PK and biodistribution studies in at
least one species will probably be needed.
8.2.8 Development of Previously Approved

or Previously Tested Drugs
The nonclinical package used to support the development of an old-

er product (i.e., an already approved product or a product for which
clinical data are available) is expected by the FDA to be consistent
with the current standards for development of a new molecular en-
tity (NME). If there are deficiencies in the nonclinical package, it is
160 S. Wilson
likely that additional nonclinical studies will be requested to address

the insufficiencies. It may, however, be possible to provide a strong
scientific justification that the studies are not needed to ensure the
safe use of the drug product if the available human data are robust
or do not identify any safety issues. Since the genotoxic potential
can not be addressed by clinical data, it will probably be necessary
to conduct the in vitro genotoxicity assays as a minimum.
8.3 U.S. Regulatory Considerations

for the Nonclinical Development of Radiotherapeutics
Development of a radiotherapeutic drug poses a number of regulatory

challenges. Because the radiotherapeutic, unlike a radiodiagnostic
drug, is intended for the palliation or treatment of a disease condi-
tion, the radiotherapeutic typically exhibits pharmacological activity.
Usually the drug exhibits efficacy even without the radionuclide being
present. The radionuclide is incorporated to potentiate the efficacy of
the unlabeled drug. Radiotherapeutics will generally be administered
for multiples doses at a larger mass dose than a radiodiagnostic. It is
common for radiotherapeutics to be used in serious and life-threaten-
ing diseases and humans are typically dosed at or near the maximum
tolerated dose. Consequently, there is an extremely narrow margin of
safety. In addition, a radiopharmaceutic may be administered by an
uncommon route (i.e., intralesional, intrahepatic artery) to maximize
exposure at the intended site, e.g., tumor.
Clearly, the FDA draft guidance: "Medical Imaging Drug and
Biological Products; Part I" (2003) does not apply to the develop-
ment of a therapeutic radiopharmaceutical and there are currently no
guidance documents that specifically address the nonclinical toxicity
study requirements for this class of drugs. If the drug to which the
radionuclide has been added is an NME, development will be con-
sistent with the ICH M3 guidance document noted earlier. Studies to
address pharmacology, safety pharmacology, pharmacokinetics, tox-
icity, genotoxicity, reproductive toxicity, and special toxicity studies
will need to be conducted to support the safety of the new drug. The
nature and scope of the nonclinical studies are comparable to those
for the development of any new drug.
If the drug to which the radionuclide has been added is an al-

ready approved drug, as is often the case, and if the route, dose, and
duration of the proposed new clinical use is the same or less than al-
ready approved, it will probably not be necessary to conduct any ad-
ditional nonclinical toxicity, safety pharmacology, or genotoxicity
studies with the cold or degraded drug substance. Biodistribution
studies conducted with radiolabeled or "hot" material, however,
would still be needed in order to calculate the radiation dose that
will be administered to humans. If the proposed new clinical use for
the drug includes higher doses, more frequent administration, longer
duration of administration, or a different route of administration and
if the current nonclinical studies do not support the proposed
changes, additional nonclinical studies will need to be conducted
with cold drug.
8.3.1 Evaluation of Potential Radiation Effects

of a Radiotherapeutic
The studies conducted with degraded or cold drug can address the
inherent toxicity of the radiotherapeutic. The FDA, however, has ad-
ditional concerns pertaining to potential radiation effects of radio-
therapeutics and more specifically to the potential for latent or late
effects that may occur in the clinical setting over 6 months to a year
after administration of the drug. As noted earlier, radiotherapeutics
are frequently used to treat life-threatening conditions. It could be
argued, therefore, that the risk of a late radiation effect is reasonable
if the patient's expected life span without treatment is very short,
i.e., less than 3-6 months. The FDA, however, has suggested that it
is important to identify the potential for late effects in order to be
able to accurately assess the risk-to-benefit ratio.
A clinical trial could theoretically be designed to assess acute radia-
tion toxicity, e.g., those effects which occur within 1-2 months of dos-
ing. A clinical study designed to allow identification of late radiation
effects prior to dose escalation (i.e., 6-18 months between doses),
however, is not feasible. Because of the concerns of delayed toxicity
effects in humans that may be associated with the radionuclide and
because the effects are often life threatening or irreversible, the FDA
162 S. Wilson
will require nonclinical toxicity studies for radiotherapeutics be con-

ducted using the hot drug. Studies with hot drug will be requested
whether the drug is an NME or an already approved product.
The "hot" studies are indicated if tissue tolerance limits are ex-
ceeded or if standard dosimetry is not applicable. While it may be
argued that dosimetry data are frequently adequate to ensure the safe
use of a radiotherapeutic and provide an appropriately predictive pa-
rameter of potential radiation damage, the FDA has indicated that
there are several scenarios where they believe that dosimetry alone
may not be adequately predictive of radiation effects. The first sce-
nario occurs when there is a mass dose effect for the drug. A mass
dose effect is evidenced by an alteration of the biodistribution of the
drug and radionuclide at higher doses than those observed with the
drug when used as a radiodiagnostic.
The second scenario is probably of greater concern and that is if
the relative biological effect (RBE) of the drug/radionuclide is great-
er than one. RBE, a measure used to compare different radiations
against a reference radiation dose, can vary considerably depending
on the organ, tissue, or endpoint that is evaluated. Assessment of
RBE can be fairly simple, but the complexity of the evaluation in-
creases as the complexity of the test system increases (Hall 2000).
The RBE may be greater than one if the drug or some component of
the clinical formulation is a radiosensitizer; if the radionuclide has
high linear energy transfer; or if there is microdistribution of the
radionuclide to sensitive targets such as DNA (i.e., Auger electron
emitters) (Howell et al. 1993).
The rationale for conducting the studies in animals with "hot"
drugs is in large part similar to the rationale for conducting any stan-
dard toxicity studies. The studies are intended to characterize the
toxicity profile of the radiotherapeutic before the drug is introduced
into humans, identify the potential target organs and safety margins,
determine the progression and potential reversibility of findings uti-
lizing histopathology, and provide a basis for establishing appropri-
ate doses for clinical use based on a risk-benefit analysis. The stud-
ies will provide data to support the most appropriate method for de-
termining dose, such as mCi based on body mass or surface area or
rads/target tissue. In addition, by using healthy animals, there are no
confounding factors that may be associated with the disease process.
8.3.2 Considerations for Conduct of a Toxicity Study

with DruglRadionuclide
To address the potential for late radiation effects and to support clinical
trials (especially repeat dose trials) with a radiotherapeutic, the FDA is
likely to request that a 1- to 2-year repeat-dose study in two species
with at least one interim sacrifice be conducted with the "hot" drug.
If the radionuclide is a fJ-emitter, then a large animal species such as
a dog, minipig, or nonhuman primate may be required. It is felt that
with fJ-emitters, use of a small animal species may overpredict the ra-
diation effects because the radiation particles can penetrate tissues for
relatively long distances. The endpoints should include standard non-
clinical toxicity parameters such as clinical observations, body
weights, ophthalmic examination, clinical pathology, gross necropsy,
organ weights, and histopathology. In addition to the long-term noncli-
nical toxicity studies, collection of dosimetry data in the species used
for toxicity studies, using two methods, will likely be requested. The
biodistribution methods should include standard direct organ and tis-
sue counts as well as the imaging modality to be used in the clinical
setting. It has been suggested that using data from the two methods
allows for the most accurate extrapolation of the animal data to hu-
mans. The actual design of the nonclinical studies to address the poten-
tial for both acute and late radiation effects, however, should be deter-
mined on a case-by-case basis. An alternate plan to that proposed by
the FDA for the nonclinical assessment of a radiotherapeutic may be
considered if scientifically sound and if it appropriately minimizes hu-
man risk. It is beneficial to obtain concurrence from the FDA prior to
initiation of the pivotal nonclinical studies in order to minimize the
possibility of additional nonclinical studies being required.
8.4 Conclusion
Nonclinical development of radiopharmaceutical drugs must take

into consideration not only the potential toxicity of the drug but the
potential radiation toxicity associated with the radionuclide. Because
radiodiagnostics are administered in typically small doses, the poten-
tial for toxicity is usually low and the potential for radiation toxicity
164 S. Wilson
is appropriately addressed by dosimetry data and limit thresholds.

FDA Guidance for Industry: "Medical Imaging Drug and Biological
Products. Part I: Conducting Safety Assessments, Draft Guidance"
(FDA 2003) provides a clear and concise outline of what is consid-
ered by the FDA as an appropriate nonclinical package for the de-
velopment of a radiodiagnostic. The nonclinical package to support
development of a radiotherapeutic will be considerably more exten-
sive than a radiodiagnostic because the potential for toxicity is con-
siderably greater with a radiotherapeutic than a radiodiagnostic. In
addition, the FDA has identified delayed or late radiation effects of
radiotherapeutics as a serious risk that should be evaluated in ani-
mals prior to exposure in humans. Consequently, nonclinical toxicity
and biodistribution studies conducted with hot drug will likely be re-
quired by the FDA to support the clinical trials in addition to studies
conducted with degraded or cold drug. Ultimately, the nonclinical
development of a radiotherapeutic should be consistent with princi-
ples that apply to the nonclinical development of any drug, but the
exact scope and type of nonclinical studies that are conducted
should be determined on a case-by-case basis.
References
Code of Federal Regulations (2002) 21 CFR 312.10 and 21 CFR 314.90

FDA (1996) Guidance for industry. Single dose acute toxicity testing for
pharmaceuticals (August 1996). U.S. Dept. of Health and Human Ser-
vices, FDA, COER, CBER
FDA (2003) Guidance for industry. Medical imaging drug and biological
products. Part I. Conducting safety assessments, draft guidance (May
2003). U.S. Dept. of Health and Human Services, FDA, COER, CBER
Hall EJ (2000) Linear energy transfer and relative biological effectiveness.
In: Radiology for the radiobiologist. Lippincott Williams and Wilkins,
Philadelphia, pp 112-123
Howell RW, Narra VR, Kandula SRS, Rao DV (1993) On the equivalent
dose for Auger electron emitters. Radiat Res 134:71-78
ICH (1997 a) Guidance for industry: M3 nonclinical safety studies for the
conduct of human clinical trials (July 1997). U.S. Dept. of Health and
Human Services, FDA, COER, CBER, ICH
ICH (1997b) Guidance for industry: S2B genotoxicity: a standard battery
for genotoxicity testing of pharmaceuticals (July 1997). U.S. Dept. of
Health and Human Services, FDA, COER, CBER, ICH
ICH (1997 c) Guidance for industry: S6 preclinical safety evaluation of bio-

technology-derived pharmaceuticals (July 1997). U.S. Dept. of Health
and Human Services, FDA, CDER, CBER, ICH
ICH (2001) Guidance for industry: S7A safety pharmacology studies for hu-
man pharmaceuticals (July 2001). U.S. Dept. of Health and Human Ser-
vices, FDA, CDER, CBER, ICH
9 The Japanese Perspective
on Radiopharmaceuticals
H. Mayahara
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

9.2 Present Status of Development of Radiopharmaceuticals
in Japan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
9.2.1 Fundamentals Relating to the Development
of New Radiopharmaceuticals . . . . . . . . . . . . . . . . . . . . . . 169
9.2.2 Number of Radiopharmaceuticals Launched Recently
in Japan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
9.3 Nonclinical Studies Necessary for the Start of Clinical Trials
of Radiopharmaceuticals . . . . . . . . . . . . . . . . . . . . . . . . . 172
9.3.1 Indications in the Guidelines . . . . . . . . . . . . . . . . . . . . . . 172
9.3.2 A Case of Consultation with Kiko . . . . . . . . . . . . . . . . . . . 172
9.4 Why Does the Japanese Regulatory Agency Often Require
More Nonclinical Studies than other Agencies Do? . . . . . . . . 173
9.4.1 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
9.4.2 The Possible Reasons Why the Japanese Regulatory Agency
Requires More Nonclinical Safety Studies Than Other Regional
Agencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
9.4.3 Scarcity of Specialists in the Regulatory Agency . . . . . . . . . 176
9.4.4 The High Level of Concern for Safety in Japan . . . . . . . . . . 178
9.5 Why is it Sometimes Difficult to Have a Scientific Discussion
with Regulatory Officials in Japan? . . . . . . . . . . . . . . . . . . 182
9.5.1 Lack of Culture of Discussion in Japan . . . . . . . . . . . . . . . . 182
9.5.2 The Remnant of Historical Social Class System in Japan ... . 183
9.5.3 Persisting Sense of Social Class System in the Present Japan .. 184
9.5.4 Examples of the Remaining Sense of the Social Class System
in Japan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
9.6 Why is the Development of New Radiopharmaceuticals
Difficult in Japan? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
168 H. Mayahara
9.7 Future of Radiopharmaceuticals in Japan . . . . . . . . . . . . . . 187

9.7.1 Advantages of Ion Beam Cancer Therapy to
Radiopharmaceutical Therapy . . . . . . . . . . . . . . . . . . . . . 187
9.7.2 Some Good News to Indicate Changes in the Regulatory
Agency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
9.1 Introduction
I am not necessarily the best person to make a presentation under the

title "The Japanese Regulatory Perspective of Radiopharmaceuticals",
since I am neither a regulatory official nor an expert of radiopharma-
ceuticals. I accepted the invitation to speak because I had heard that
Schering Japan Ltd. had encountered great difficulty in finding a
speaker, and I had little doubt that no Japanese regulatory officials
or ex-regulatory officials would come to such a meeting and speak un-
der the title of "Regulatory Implications" and "Radiopharmaceuticals".
I had, however, some reasons to accept the invitation as follows.
First, I was once a government official of the Ministry of Education,
having been a university teacher for 16 years, 11 of the 16 years in
a national university. Second, after moving to Takeda Chemical Indus-
tries, the largest pharmaceutical company in Japan, I had a long experi-
ence of working with government officials. During the 20 years of ser-
ving the company, I worked for Japan Pharmaceutical Manufacturers
Association (JPMA) , joined the ICH (International Conference on
Harmonisation of Technical Requirements for Registration of Pharma-
ceuticals for Human Use) activities and worked together with Japanese
regulatory officials to develop various nonc1inical ICH guidelines for
7 years. Third, I was the rapporteur of the ICH Topic M3, "Timing
of Nonc1inical Safety Studies for the Conduct of Clinical Trials" from
its start to the completion of the guideline, and became familiar with
the regional differences in the nonc1inical safety studies. The fourth
reason was a personal one: one of my sons happened to be a radiation
oncologist, and I could count on his help to obtain information on
radiopharmaceuticals and cancer therapy using them.
The reason why no Japanese regulatory officials will come to
such a meeting is simple: the Japanese government prohibits offi-
cials from having side jobs (even when they use their holidays). In
The Japanese Perspective on Radiopharmaceuticals 169
addition, the government also prohibits officials from having perso-

nal contacts with private corporations. Semi-government officials
such as personnel of The Organization for Pharmaceutical Safety
and Research (OPSR, Japan; Kiko), an organization in charge of
consultation at the Investigational New Drug (IND) and New Drug
Application (NDA) levels, and of good laboratory practice (GLP)
and good clinical practice (GCP) inspections in Japan, should be-
have just the same as government officials.
Government officials can attend nonprofit or academic meetings.
However, if the name of the meeting is crowned by a private com-
pany, it may be sometimes difficult to be recognized as a nonprofit
meeting in Japan. Further, in the case where government officials
could attend the meeting, another difficulty is the scarcity of regula-
tory scientists in the Japanese government. The reasons for this scar-
city are discussed later.
9.2 Present Status of Development

of Radiopharmaceuticals in Japan
9.2.1 Fundamentals Relating to the Development

of New Radiopharmaceuticals
At first, fundamentals to support the development of new radiophar-

maceuticals should briefly be surveyed. Figure I shows the compari-
son in the number of X-ray computed tomography (CT) and mag-
netic resonance imaging (MRI) machines per million people among
developed countries (OECD 2000). As shown in the figure, the num-
bers of X-ray CT and MRI machines per capita in Japan are ex-
tremely high compared with those in other developed countries. This
leads to the conclusion that Japanese doctors or hospitals conduct
more diagnostic tests than those in other countries.
Why so Japanese doctors or hospitals conduct more diagnostic
tests than those in other countries? This tendency could be explained
as originating from the following situations:
1. In Japan, the prices of drugs, medical treatments and diagnostic

tests are determined by the government and listed in The List of
170 H. Mayahara
U.S.A UK Germany France Sweden Japan

Fig. 1. Number of X-ray CT and MRI machines in the USA, UK, Germany,
France, Sweden and Japan per million population. Note the abnormally high
numbers in Japan
National Health Insurance Price Standard, and medical expenses

are calculated by a pay-for-service system. This means that the
more prescriptions and diagnostic tests are conducted, the greater
the incomes will be for doctors and hospitals.
2. The medical treatment fees in The List of National Health Insur-
ance Price Standard are the same for all doctors regardless of
their expertise and years of experience except for some specified
highly advanced treatments. This means that Japanese doctors and
hospitals tend to lean toward over-prescription and excessive diag-
nostic tests when they want to increase their incomes.
3. Recently, the government has introduced a policy of separating
hospitals from pharmacies to decrease national medical care ex-
penditures. The introduction of this policy has succeeded in de-
creasing over-prescription. The government has also introduced a
partial blanket medical expense system; however, this has not yet
succeeded in decreasing the number of diagnostic tests.
Judging from the abundance of X-ray CTs and MRIs in Japanese hos-
pitals, you might assume that the fundamentals supporting the devel-
opment of radiopharmaceuticals exist in Japan, and the development
of radiopharmaceuticals is easy. However, this is not true. The number
of gamma cameras, positron emission tomography (PET), PET-CT and
single photon emission computed tomography (SPECT) scanners,

which are more directly related to the development of new radiophar-
maceuticals, is surprisingly few in Japan. For example, there exists
only one PET-CT, a scanner used for therapeutic purposes (Hisada
2002). The reason why these instruments are poorly represented in Ja-
pan is discussed later.
9.2.2 Number of Radiopharmaceuticals Launched Recently

in Japan
Figure 2 shows the annual number of radiopharmaceuticals launched

in Japan in the past 30 years (Pharmaceutical Evaluation Study
Group 2001a). As is evident, 53 radiopharmaceutical drugs have
been approved for marketing. However, approval of radiopharmaceu-
ticals for therapeutic use number only two. These both are drugs de-
rived from the same substance, radioactive iodine, and their approval
was 13 years ago.
Another thing to be read from Fig. 2 is that the approval of radio-
pharmaceuticals has apparently decreased in the last 10 years. These
data clearly show that approval of radiopharmaceuticals for thera-
peutic use is difficult in Japan. The possible reasons why the ap-
• Therapeutics
--- -
~- • Diagnostics
-
II I I II I
15 90 DII
Fig. 2. The annual number of radiopharmaceuticals launched in Japan in the

past 30 years. Note that among 53 radiopharmaceuticals approved in Japan
in the last 30 years, only two have been approved for therapeutic use, and
they both are drugs of the same substance, radioactive iodine
172 H. Mayahara
proval of radiophannaceuticals is difficult in Japan will be discussed

in the Sect. 9.6.
9.3 Nonclinical Studies Necessary for the Start

of Clinical Trials of Radiopharmaceuticals
9.3.1 Indications in the Guidelines
Regular nonclinical studies described in the ICH Topic M3 guideline

(ICH 1997) may be considered at the start of clinical trials for radio-
pharmaceuticals in Japan. Judging from our limited experiences of
consultation with Kiko, no Japan-specific additional animal studies
seem to be required for the start of clinical trials in Japan if foreign
clinical-trial data are available. However, in the case of staging up
of clinical trials and of the marketing approval, Japan-specific non-
clinical studies are often required as discussed later.
Concerning the nonclinical studies necessary for the submission
of radiophannaceuticals in Japan, no guidance or guidelines are
available except for the "SEIZOU-SHISHIN", Pharmaceutical Manu-
facturing Guidelines (Pharmaceutical Evaluation Study Group
2001 b). They only mention: "The biological effect of radiopharma-
ceuticals has generally been understood mainly as that of the radio-
isotope contained. Therefore, if there is a scientifically proper rea-
sones) for skipping animal studies the document concerned is not re-
quested. The necessity of the animal study should be considered
adequately by the applicant."
In addition, the carcinogenicity study guideline (MHLW 2003)
describes: "These studies are usually not necessary for radiopharma-
ceuticals."
9.3.2 A Case of Consultation with Kiko
Since there are no specific guidelines or guidance for the nonclinical

development of radiopharmaceuticals, we have to make the best use
of consultation with Kiko in order to promote the development of
new radiophannaceuticals efficiently.
A case of Company A in consultation with Kiko during the devel-

opment phase of a radiopharmaceutical is of interest to consider the
fate of a drug candidate in this category. In a consultation meeting
concerning the development of a drug candidate which was an iso-
tope-labeled monoclonal antibody against a human protein, Kiko re-
quired Company A to conduct animal safety studies using the final
drug product, designed for human use. The company took exception
with that stance, stating that the effects of monoclonal antibody
against a human protein on normal animals would be completely dif-
ferent from those on human patients, and they did not think any
meaningful information would be obtained from the requested stud-
ies. The company also stated that the FDA had already approved the
drug without requiring such animal studies, and they could submit
appreciable amounts of human data. However, the official said,
"This is a crush of viewpoints. We want to evaluate more animal
data", and the official refused further discussion.
The following three questions could be derived from the above
episode and other similar reported episodes of other pharmaceutical
companies:
l. Why does the Japanese regulatory agency often require more non-
clinical studies than other agencies do?
2. Why is it sometimes difficult to have scientific discussions with
regulatory officials in Japan?
3. Why is the development of new radiopharmaceuticals difficult in
Japan?
9.4 Why Does the Japanese Regulatory Agency Often

Require More Nonclinical Studies than other Agencies Do?
9.4.1 Examples
Historically, the Japanese regulatory agency had a tendency to re-

quire more nonclinical safety studies than other regional agencies.
This tendency has been shown not only at the start of the first hu-
man clinical trials, but also at the stepping up of the clinical devel-
opment stages or at the marketing approval (Scales and Mahoney
174 H. Mayahara
1991; Griffin 1992; Parkinson 1994). This tendency has been gradu-
ally less prominent in the latest 10 years, keeping in step with the
activities of the ICH, in which more than 40 nonclinical and clinical
guidelines were completed. Especially, the ICH Topic M3 guideline,
Timing of Nonclinical Safety Studies for the Conduct of Clinical
Trials (lCH 1997), was most effective in decreasing the number of
nonclinical safety studies required in Japan at the start of the first
clinical trials by harmonizing the regulatory requirements among
three regions.
However, there are still remaining regional differences in the ICH
Topic M3 Guideline, and these are good examples showing the ten-
dency of the Japanese regulatory agency; it still requires more non-
clinical studies than other regional regulatory agencies as follows:
1. Single dose toxicity studies
Nonrodent single dose toxicity studies are still necessary in Ja-
pan.
2. Reproductive toxicity studies
Until recently and only in Japan, a rodent male reproductive tox-
icity study had been required before the start of phase I studies
with healthy male volunteers. The dosing period necessary for
this male reproductive toxicity study was 9 weeks at first (Ta-
kayama et al. 1995). To decrease the dosing period, a collabora-
tion study was conducted by JPMA and the Ministry of Health
and Welfare (MHW, at the time, now known as MHLW). In this
study, the effects of dosing for 4 weeks and 9 weeks were com-
pared using various substances known to be toxic to the reproduc-
tive organs (Takayama et al. 1995). This collaboration study
showed that detailed histopathologic examination of the male re-
productive organs after a 4-week dosing period is more sensitive
than the conventional male reproductive toxicity studies with a
dosing period of 9 weeks for evaluation of reproductive toxicity.
Based on the results of this collaboration study, detailed histo-
pathologic examination of male reproductive organs was accepted
as a method for evaluating male reproductive toxicity before start-
ing phase I studies in the Topic M3 guideline completed in 1997
(ICH 1997).
In the original ICH Topic M3 guideline, however, the minimum
dosing period required to evaluate male reproductive organ toxic i-
ty in rats was at least 4 weeks in Japan and 2 weeks for the USA
and ED. As the reason for the longer requirement, Japanese
authorities explained that there were not enough data to show that
a 2-week dosing period is sufficient to evaluate male reproductive
toxicity (ICH 1997).
To dissolve these discrepancies among three regions, a similar
collaboration study was conducted by JPMA and MHW to com-
pare the effects of dosing for 2 weeks and 4 weeks using various
substances known to be toxic to the reproductive organs (Sakai et
al. 2000). The results of this collaboration study showed that 2-
week dosing is as effective as 4-week in histopathologic evalua-
tion of male reproductive organ toxicity. Thus the ICH Topic M3
guideline was revised and the minimum duration for the evalua-
tion of male reproductive toxicity in rats to support phase I and II
studies up to 2 weeks was harmonized to 2 weeks in 3 regions in
2002.
An intrinsic male reproductive toxicity study is to be conducted
before the start of phase III studies in the ICH Topic M3 guide-
line (ICH 1997). However, the Japanese regulatory agency still re-
commends the study to be conducted before the start of the pha-
se I study. The note of the revised reproductive toxicity guideline
says, "In Japan, male reproductive toxicity studies in rats are
usually conducted before the start of phase I clinical trials using
healthy male volunteers" (MHLW Pharmaceutical and Food
Safety Bureau, Evaluation and Licensing Division 2000).
Another example of heavier requirements in reproductive toxicity
studies in Japan is that female reproductive toxicity studies are
necessary when women of childbearing potential participate in
phase I trials, regardless of their contraceptive conditions.
An additional example in reproductive toxicity studies in Japan is
that reproductive toxicity studies, even for submission of cyto-
toxic anticancer drugs, are usually necessary.
3. Antigenicity study and drug dependency study
These studies are usually required in Japan even for drugs with
abundant human use in foreign countries unless there is firm evi-
dence that these studies are not necessary.
176 H. Mayahara
9.4.2 The Possihle Reasons Why the Japanese Regulatory

Agency Requires More Nonclinical Safety Studies
than Other Regional Agencies
There are two possible answers for this question. The first is the
scarcity of specialists of new drug development in the regulatory
agency, and the second is the generally higher degree of concern for
safety in Japanese society than in other regions, as explained below.
9.4.3 Scarcity of Specialists in the Regulatory Agency
In the Japanese government, there are very few specialists who have
the experience of new drug development in pharmaceutical compa-
nies. If a regulatory official has no experience of new drug develop-
ment, he may have little specialist knowledge on new drug develop-
ment and perhaps no cost or efficiency consciousness that company
employees usually have.
When a pharmaceutical company proposes to do without some
safety studies at a consultation with Kiko, there are at least three
possible reasons:
1. The study is needless for scientific reasons.

2. Its necessity is scientifically doubtful or very low.
3. Its scientific necessity cannot be denied, but it will cost a lot of
money or time.
Therefore, if a regulatory official does not have the requisite special-

ist knowledge or sense of business or cost, the official may have dif-
ficulty in understanding the true reason(s) why a pharmaceutical
company proposes to omit some safety studies. If he cannot find the
true reason(s) to skip some safety studies, he may regard the propo-
sition as coming merely from saving money, seeking for profit and/
or neglect of safety for the reasons discussed later in this chapter.
Thus, the official will have no hesitation refusing such proposals and
requiring all the safety studies he deemed necessary, regardless of
the financial or time commitments, or degree of necessity, resulting
in the increase in the requirements for noncIinical studies.
The above explanation may produce another question: Why are

there very few regulatory officials in the government who have ex-
perience working for private firms in Japan?
9.4.3.1 The Reason Why There Are Very Few Regulatory

Scientists in the Japanese Government Who Have Experience
Working in Private Companies
Under the Japanese civil-service law, almost all government officials are
selected using the civil-service examination at college graduation, ex-
cept for national university teachers. Such officials usually stay in the
government until their retirement. This means that they become govern-
ment officials without any experience in the research or new drug de-
velopment business. In addition, the government prohibits officials from
working for private companies or even to have personal contacts with
private companies. This means that they have little chance of gleaning
the latest know ledge in new drug development or gaining a sense of the
business, cost and efficiency directly from private companies.
Such a situation is in clear contrast with that in the U.S. In an
ICH meeting, one of the FDA representatives showed me a card that
he always carried in his purse. The card was a license allowing him
to work for private companies within certain limits per week. He ex-
plained that the purpose of this FDA policy to permit officials to
work for private companies was the on-the-job training COJT) for of-
ficials. With this system, the FDA officials can obtain the latest
knowledge on new drug development and can keep their expertise
and sense of reality in handling issues relevant to business in the
pharmaceutical industry. If you understand the above differences in
the situations of government officials between Japan and the U.S.,
you will be able to understand the reason for the different reactions
of the Japanese government- or semi-government officials versus
those of other regions at consultation meetings.
9.4.3.2 The Reason Why the Government Does Not Employ

Enough Officials or Specialists from Private Companies in Japan
You may ask why the Japanese government does not employ enough
specialists if they are so necessary.
Historically, there was no custom to recruit the government offi-
cials from private companies, though the reason has never been offi-
178 H. Mayahara
cially explained. Also, government officials rarely move to private

companies in Japan. In other words, mobility of the employees be-
tween the government and private companies is almost negligible in
Japan, though this custom is showing a minor change recently.
One of the possible reasons why the government rarely employs
officials from private companies is the restriction of the civil-service
law which stipulates that the government officials should be selected
using the civil-service examination.
Another explanation for the low mobility of personnel between
the government and private companies in Japan is the difficulty in
changing the civil-service law. Japan is a very conservative country,
especially in changing the law. For instance, Japan has never changed
even a single word in the constitutional law after its promulgation in
1946, in contrast to Germany, which has revised its constitution 38
times.
Apparently, the above two explanations do not sufficiently explain
the reason why the Japanese government limits the method for recruit-
ing officials to the winners of the civil-service examination and avoids
recruiting them from private companies. The above explanations also
do not explain the fact that, although a new gate was opened recently
for workers of private companies to become government officials by a
special legislation, the number of officials employed through this new
route is very few. Therefore, it is apparent that a different explanation
is necessary to understand the reason why only few government offi-
cials are employed from private companies in Japan. I will discuss
the additional explanation later this chapter.
9.4.4 The High Level of Concern for Safety in Japan
As explained earlier, the hypothesis to explain the reason for the ten-
dency of the Japanese regulatory agency to require more nonc1inical
safety studies than other agencies is to regard that it is merely the
reflection of the relatively high concern for safety and health in Ja-
pan. There are at least two reasons for this high concern for safety
in Japanese society: the first is the effect of the defeat of Japan in
the Second World War, and the second is the existing higher degree
of social safety in Japan as explained below.
9.4.4.1 Effects of the Defeat in the World War II

During the Second World War, over 3 million Japanese soldiers and
citizens were killed, including more than 300,000 victims of the two
atomic bombs. After the war, peace and safety became the national
norm in Japan. The seeking after peace and safety in Japan seems to
have been carried to excess as follows:
- The Japanese constitution prohibits this country to have armed

forces.
- The self-defense army was founded during the Korean War
against the fear of socialist countries, but the constitution has not
been revised.
- The self-defense army cannot have weapons that may threaten
neighboring countries, such as aircraft carriers, long distance mis-
siles and nuclear weapons.
- The PKO corps dispatched from Japan is prohibited from carrying
powerful weapons and operating in combat zones.
- At hijackings, the Japanese government usually seeks negotiated
settlements with ransom money.
- The Japanese government often prohibits personal adventure, such
as crossing the Pacific Ocean single handedly in a sailing boat,
because it is dangerous.
- Some local governments prohibit the climbing of some specified
mountains through some specified routes because they are too
dangerous.
The seeking for safety in Japan has not been limited to physical
safety but has been expanded to the economic safety as follows:
- Private gambles are prohibited in Japan.

- Casinos are prohibited in Japan.
- Fundamental rights of workers have been protected in Japan by
social customs such as the lifetime employment, promotion by se-
niority and cooperative intracompany labor unions, though these
good old customs are changing recently due to globalization.
- The social gap in income levels has been controlled to minimum
with heavy progressive taxation.
180 H. Mayahara
The social security system has covered all the poor, aged and
handicapped.
- Universal coverage of public health insurance and low official
medical costs to all Japanese was realized in 1961, which helped
lead to the longest life expectancy and the lowest death rate of in-
fants in the world.
9.4.4.2 The Reason Why Japan is a Very Safe Country

The reason why Japan is such a safe country is partly because of the
nation's keen seeking after peace and safety after the defeat in the
Second World War, as stated in the previous section. However, his-
torically Japan has always been a safe country, mainly due to geo-
graphical and historical reasons.
- Being isolated in the Pacific Ocean, Japan was rarely attacked by

foreign armed forces, and has never been occupied by a foreign
government, except for the temporal occupation by the U.S. after
the defeat in the Second World War.
- Historically, there have been almost no racial or cultural conflic-
tions in Japan, because Japan has been almost mono-race, mono-
lingual and monoculture country.
- Historically, there has been almost no religious conflict in Japan
because most Japanese are polytheists.
- Japan was the first country to realize an armless society with the
"sword hunt" of 1588. Even at present, the rate of murder in Ja-
pan is less than l/1O that of the U.S. (6.8 for the U.S. and 0.62
for Japan per 100,000 population), and the rate of murder with
guns is 230 times less frequent in Japan than that in U.S. (4.61
for the U.S. and 0.02 for Japan) (GunCite 2003).
- As murders with guns are so rare in Japan (0.07/day in contrast
to 32.6/day in the U.S.), they will be big news when they occur.
- Firing of guns itself becomes big news in Japan. Even when a po-
lice officer uses his handgun against a criminal and in self-de-
fense, the number of discharged bullets, the number and the loca-
tions of hits are reported by radio, TV and newspapers with a na-
tionwide circulation with commentary on possible justifications
for use of the gun.
- The unemployment rate, divorce rate and crime rate per popula-
tion in Japan are the lowest in the world, although these indices
are increasing recently.
- The numbers of lawyers and court cases are very low in Japan.
The approximate number of lawyers is 900,000 in the U.S., but
only 18,000 in Japan, although the population of Japan is about
one half of that of the U.S.
9.4.4.3 High Level of Concern for Factors That Might Threaten

Safety in Japan
The high level safety in Japanese society results in a high level of
concern for factors that may threaten safety or health. There is a lot
of circumstantial evidence showing the high level of concern for
safety or health in Japanese society:
- The flood of antibacterial goods in drug stores and in stationery

shops
- The flood of health drinks (supplement drinks) in drug stores,
kiosks and food shops
- The volume of natural foods, nutrient-enriched foods and agricul-
tural-drug-free foods available
- Emotional rejection of genetically modified foods
- Too much anxiety over drug-adverse effects
- Hesitation towards using high-dose narcotic drugs, even for pallia-
tive purposes
- Rejection of artificial radioactive substances and radiopharmaceu-
ticals
- Difficulty in finding volunteers for clinical trials.
The cause for the heavier requirement for nonclinical safety studies
in Japan versus other regions may be a reflection of the above-men-
tioned high level concern for safety and health among Japanese peo-
ple.
182 H. Mayahara
9.5 Why is it Sometimes Difficult to Have

a Scientific Discussion with Regulatory Officials in Japan?
There are several reasons why it is sometimes difficult for pharma-

ceutical company workers to have scientific discussion with regula-
tory officials in Japan. As stated in Sect. 9.4.3, the scarcity of spe-
cialist and specialist knowledge on new drug development in regula-
tory officials may partly be responsible for the difficulty in having
scientific discussions. However, there are at least two other reasons,
which may be difficult for foreigners to understand. One is the lack
of culture of discussion in Japan and another is the remnant of his-
torical social class system in Japan.
9.5.1 Lack of Culture of Discussion in Japan
The main reason why scientific discussion with regulatory officials

is sometimes difficult in Japan is the "lack of culture of discussion"
or the Japanese culture not to discuss. There are at least two factors
which support this "lack of discussion" in Japan. The first is the cul-
ture of "wordless understanding" in Japan and the second is honori-
fic expressions in Japanese language.
9.5.1.1 'Wordless Understanding' in Japan

It has long been believed in Japan that the best way of understand-
ing each other is the heart-to-heart communications using no words.
In other words, the quality of communication by language has long
been believed to be far inferior to the heart-to-heart communica-
tions. For instance, the typical Japanese husband and wife almost
never say to each other "I love you". Without such communication
by words, the divorce rate in Japan is far below those in the U.S.
and European countries. This culture of "wordless understanding"
originates from the almost mono-race, monolingual and monoculture
situations unique to Japan. However, this culture has been an obsta-
cle to developing the culture to discuss or debate in Japan.
9.5.1.2 Honorific Expressions in Japan

Every Japanese sentence in conversation must contain honorific ex-
pressions when the person you are speaking to is older than you or
has a higher professional class than your own. Suppose that you are
having conversation with someone and only you have to use very
polite expressions and the person you are talking to is not, scientific
discussion or free debate on equal footing is very difficult. Perhaps
this is the main reason why the culture to discuss or to debate has
never developed in the history of Japan. The introduction of democ-
racy and U.S. culture after the defeat of Japan in the Second World
War have not much changed the construction of Japanese society
and the construction of the Japanese language. Thus in Japan, scien-
tific discussion is difficult not only in the consultation meetings be-
tween the regulatory officials and company workers but also among
the scientists in academic meetings.
The reason why the consultation with Kiko often results in a
crush of viewpoints can be partly explained with the above hypoth-
esis, "lack of culture of discussion".
9.5.2 The Remnant of Historical Social Class System in Japan
Japan is a democratic country now. However, there is still a remain-

ing sense of the social class system. To understand this, you need
some knowledge of Japanese history and the unspoken social preju-
dice present in Japan.
In the history of Japan, the regimes governed by "samurai" con-

tinued for about 800 years.
- The "samurai" were expected to discard their lives for their lords
or for justice whenever it was necessary.
To protect the samurai's norm from moral corruption, the "samurai"
were kept always poor and expected not to seek for worldly profit.
Norm for the "samurai" became the social norm in Japan in prin-
ciple, because the samurai regimes continued so long.
- The order of social classes, "samurai - farmer - craft worker -
merchant", was established during the 265 years of the Edo re-
gime (1603-1868).
184 H. Mayahara
It is important to note that the merchants, the richest class people,

were ranked to the lowest in this class system. It is important to
note that the sense of social class system still remains in the depth
of the heart of the Japanese, even at the present day, and this sense
of social class may partly be responsible for the difficulties in hav-
ing scientific discussions between the regulatory officials and private
company workers on an equal footing as stated below.
9.5.3 Persisting Sense of Social Class System in the Present Japan
At present, the professions that may fall under the category of the
"samurai class" are government officials. Medical doctors and uni-
versity professors may also be included in this class. All these pro-
fessions are regarded as nonprofit professions.
As government officials are selected through very difficult civil-
service examinations, they are regarded as the best and the brightest
in Japan. They are proud of being one of the members of this elite
class, and usually have, in the depth of their mind, a sense of super-
iority to the private company workers that are engaged in profit-
making businesses. This sense of superiority in the government offi-
cials may explain the difficulty in having discussion with company
workers on an equal footing.
The hypothesis that there exists elitism in government officials
also effectively explains the fore-mentioned custom that the Japa-
nese government almost never employs officials from private compa-
nies. This is because the government may regard the workers in the
private companies as if they have lost the qualification to be govern-
ment officials, because they once had worked in private companies
and therefore they would have been "contaminated by the profit-first
way of thinking" of the lowest class of people.
9.5.4 Examples of the Remaining Sense

of the Social Class System in Japan
The following two episodes are examples showing the remammg

sense of the social class system in the people belonging to the "sa-
murai class" in the present time.
The first example is a newspaper article which I found a few
years ago. The president of the Japan Medical Association, a famous
medical doctor, said that to decrease national medical expense, it is
wrong to cut down the doctor's fee because doctors are not paid
enough. He said that instead the national official drug prices should
be cut down because there is a pharmaceutical company that has
made annual profit of more than one billion dollars. Perhaps he said
this because he felt that making profit by selling medicines for poor
patients is almost equal to committing a crime. However, it is very
strange that he pays no consideration to the fact that Japan is not a
socialist country and that pharmaceutical companies in Japan are
stock companies. Japanese pharmaceutical companies not only have
to remain profitable and share profits with stockholders, but also
they have to deal with the foreign-capital, giant pharmaceutical com-
panies which are working their way into the Japanese market in this
internationalized era. It is very strange that a highly educated, so-
cially recognized, famous medical doctor voices such an opinion, it
being so far from the reality. This episode indicates his complete
lack of a sense of business and economics and/or his sense of supe-
riority towards the workers in pharmaceutical companies.
The second example is the disclosure of the so-called official
ranking list of billionaires by the Japanese government. In Japan, the
"Official Ranking List of Largest Taxpayers" is disclosed once every
year by the National Tax Administration Agency on newspapers and
TVs. One can calculate the annual income of billionaires easily
from the amount of income tax disclosed. However, what is the pur-
pose of this disclosure? Perhaps the purpose may be explained, in
principle, as a system to thank the rich people for their great contri-
butions to the national income. However, do all the rich people want
their annual income disclosed? Is it not a violation of their privacy?
Perhaps this system is based on a national prejudice that the rich are
but the lowest class people and they will cheat the government tax
186 H. Mayahara
office if their names and annual incomes are not disclosed, and no
one is watching them not to do evil.
9.6 Why is the Development

of New Radiopharmaceuticals Difficult in Japan?
As discussed in Sect. 9.2.2, the only radiopharmaceuticals for thera-

peutic use launched in Japan in the past 30 years was radioactive io-
dine, approved 13 years ago (Fig. 2). The figure also shows that the
approval of radiopharmaceuticals has apparently decreased in the
last 10 years. Not a few articles have reported the difficulty in the
approval of new radiopharmaceuticals, especially for therapeutic use,
and the decrease in the number of young doctors engaging in nucle-
ar medicine (Kimura 2000; Kusakabe 2001; Tajiri 2003). For exam-
ple, the case of radioactive 89Sr, a popular drug in other countries as
pain palliative for bone cancer, illustrates this point. Although it was
submitted for application in 1999, its approval has been suspended
(Kimura 2000). There are at least six such radiopharmaceutical can-
didates that are suspended for marketing approval (Kusakabe 200 I).
It seems that there are the following three difficulties in conducting
radiotherapy in Japan:
1. The first difficulty is the high cost of investment required to con-

duct radiotherapy. The radiopharmaceutical treatment incurs high
costs for investment for facilities and operating costs to comply
with numerous laws and regulations to handle radioactive sub-
stance.
2. The second difficulty is the low price of radiopharmaceuticals in
Japan. l3l I is very cheep on the List of National Health Insurance
Standard Prices.
3. The third difficulty is the deficiency in the National Standard
Medical Fee Table. As the fee for treatment with 131 1 is not regis-
tered on the Standard Table, it is not covered by insurance, and
under Japanese law - to prohibit the mixture of insurance-covered
and noncovered therapies - doctors cannot charge the treatment
fee for 131 1 therapy to patients. For these reasons, the radiophar-
maceutical treatment does not pay in common hospitals. The
radiopharmaceutical therapy is only possible at facilities with lit-

tle concern for profitability, such as national hospitals, or in the
very few private hospitals specialized for this purpose (Tajiri
2003).
As the base of these difficulties in approval and use of radiopharma-

ceuticals, all the Japanese will agree to point out that there is a
strong aversion to radioactivity and radiopharmaceuticals in the com-
mon Japanese. This antipathy, unique to the Japanese, comes from
the trauma of the tragedy of two atomic bombs. This strong allergy
to radioactivity results in the low number of patients who want to
have radiotherapy or medication with radiopharmaceuticals. The low
number of patients results in the low incentive for doctors and hospi-
tals to invest in radiation therapy and the decrease in the number of
doctors engaging in the nuclear medicine. It is apparent that the low
number of approved radiopharmaceuticals, slow review, delay or de
facto refusal of approval of radiopharmaceuticals by the Japanese
regulatory agency are simple reflection of the national allergy to
radiopharmaceuticals.
9.7 Future of Radiopharmaceuticals in Japan
9.7.1 Advantages of Ion Beam Cancer Therapy

to Radiopharmaceutical Therapy
As methods for cancer therapy, dissection, chemotherapy and radia-

tion therapy are popular in Japan and radiopharmaceutical therapy
faces many obstacles as described in Sect. 9.6. Recently, a new
method for cancer therapy, the ion beam therapy, is attracting atten-
tion in Japan. For instance, a facility named HYOGO Ion Beam
Medical Center (HIBMC) began cancer therapy on I April 1993.
HIBMC is the first facility in the world that has both proton and car-
bon beam facilities for cancer therapy. HIMBC is located close to
the Spring-8, the world's most powerful third generation synchrotron
radiation (SR) facility, located at Harima Research Park near Kobe,
Japan. This facility produces high-energy photons with an 8 GeV
storage ring. The SR is about 100 million times more brilliant than
188 H. Mayahara
conventional X-rays, and it is hoped that its spatial and contrast re-
solution may enable radiation oncologists in the HIMBC to visualize
very fine structures of the tumor-bearing areas and to determine the
target volume with precision. Thus, SR of the Spring-8 will allow
the HIMBC to deliver ion beam radiotherapy with unparalleled accu-
racy to the target tumor volumes (Hyogo Ion Beam Medical Center
2003; Kagawa et al. 2002).
The proton and carbon-ion beams are far superior to X- or gam-
ma-rays in locating the radiation dose to the target tumor volume. In
addition, carbon ions are nearly two times more effective than are
X- or gamma-rays in killing the cancer cells. It is therefore expected
that those ion beams can cure early stage tumors with minimal dam-
age to normal tissues.
The ion beam treatment has also advantages to the cancer treat-
ment by radiopharmaceuticals as follows:
1. The ion beam causes no pollution of the body and environment

by radioactive substances as radiopharmaceuticals.
2. Ion beam treatment has no fear of unexpected exposure for per-
sons engaged in the manufacture, transport, storage, use and dis-
position of the technology, as radiopharmaceuticals have.
3. In the case of high-dose treatment with radiopharmaceuticals, pa-
tients must be kept in a special room until the radioactivity
emitted from their bodies and their excrement is proved to be un-
der the specified standard values.
4. The ion beam treatment causes the minimal damage to the nor-
mal other tissues of the patients. Thus, patients can not only en-
joy high QOL but also expect a quick return to cancer-free nor-
mal lives and/or work (Hyogo Ion Beam Medical Center 2003).
Judging from the difficulties in obtaining marketing approval and in

therapeutic use of radiopharmaceuticals, it is apparent that ion beam
treatment will be preferred to radiopharmaceutical treatment as the
method for cancer therapy in Japan in the future.
This pessimistic estimation for the future of radiopharmaceuticals,
however, is only for their therapeutic use. The diagnostic use of
radiopharmaceuticals will increase for both conventional diagnosis
and for diagnosis before cancer therapy.
9.7.2 Some Good News to Indicate Changes in the Regulatory

Agency
There is some good news for pharmaceutical companies showing

possible changes in the Japanese regulatory agency. Table 1 shows
the number of monoclonal antibody drugs launched in the last
12 years in Japan. The uppermost one is a drug used after renal
transplantation, and its approval 12 years ago was an exception.
After its approval, no monoclonal antibody drugs were approved for
roughly 10 years. However, three monoclonal antibody drugs for
four efficacies have been approved in the last 3 years, indicating that
Table 1. The monoclonal antibody drugs approved for marketing in Japan
Trade name Type of AB Effective for Company Launched
Orthoclone Anti-CD3 Acute Yansen- May

OKT (muro- monoclonal regeneration Kyowa 1991
monab-CD3) AB after a renal Hakkou
transplantation
Herceptin Anti-HER Metastatic Nihon Roche June
(trastuzumab: human mammary 2001
genetical monoclonal cancer
recombination) AB
Rituxan Anti-CD20 Malignant Zenyaku- April
(rituximab: monoclonal B celllym- Nihon Roche 2002
genetical AB phoma
recombination)
Simulect Anti-CD25 Acute Tanabe April
(basiliximab: (IL-2 receptor) rejection after 2002
genetical monoclonal a renal
recombination) AB transplantation
Simulect Anti-CD25 Chronic Tanabe May
(basiliximab: (IL-2 receptor) arthritic rheu- 2003
genetical monoclonal matism (add.
recombination) AB indication)
The approval of Orthoclone OKT was made 12 years ago, an exceptional

move at the time. After its approval, no monoclonal antibody drugs were ap-
proved for 10 years. However, three monoclonal antibody drugs have been
approved in the last 3 years.
190 H. Mayahara
the high hurdle for approval of monoclonal antibody drugs has been
lowered.
Other good news for pharmaceutical companies is the approval of
IRES SA, an epidermal growth factor (EGF) receptor tyrosine kinase
inhibitor. This is a molecular-targeted medicine for treatment of lung
cancer developed by AstraZeneca. MHLW approved it in July 2002
for the first time in the world with a review period of only
5 months. IRES SA, however, caused death by interstitial pneumonia
for 124 patients in 5 months after its marketing. However, MHLW
has succeeded in decreasing the case of death by 1/5 after warning
for cautious administration with frequent clinical tests (Yahoo Japan
News 2002). The FDA approved IRES SA on 5 May 2003, after
prolonging the review period twice to watch the whole circum-
stances in Japan.
Historically, approval of new chemical entities (NCEs) in Japan
has seldom preceded that by FDA. Therefore, the approval of IRES-
SA by MHLW preceding that by FDA may have shown the foresight
of MHLW and/or a change of policy of MHLW to approve NCEs ac-
tively. Thus, we may have some hope that the attitude of the Japa-
nese regulatory agency surrounding the approval and use of radio-
pharmaceuticals may change in the near future.
References
Griffin JP (1992) A view from the industry. The way forward, proceedings
of the ICH-l. Greystone Books, Antrim, pp 1-16
GunCite (2003) International homicide comparison (online, cited 1 Sept
2003), http://www.guncite.comlgun_control....gcgvinco.html
Hisada K (2002) Kantougen (preface), Nippon Acta Radiol 62:461-462 (in
Japanese)
Hyogo Ion Beam Medical Center (HlliMC) (2003) http://www.
hibmc.shingu.hyogo.jp/englishlaisatu-e_top.htm
ICH (1997) Topic M3 guideline, timing of non-clinical safety studies for the
conduct of clinical trials
Kagawa K, Murakami M, Hishikawa Y et al (2002) Preclinical biological as-
sessment of proton and carbon ion beams at Hyogo Ion Beam Medical
Center. Int J Radiat Oncol BioI Phys 54:928-938
Kimura R (2000) Sr-89 bone metastasis pain relief therapy (in Japanese).
Jpn J Radiol Technol 56:885-889
Kusakabe K (2001) Present status and problems in the use of radiopharma-

ceuticals (in Japanese). Kakuigaku (Nucl Med) 38:507
MHLW Pharmaceutical and Food Safety Bureau, Evaluation and Licensing
Division (2000) On the revision of guideline for non-clinical safety stud-
ies for the conduct of human clinical trials for pharmaceuticals (in Japa-
nese). Iyakushin (Division Notice) No.1832, 27 Dec 2000
MHLW (2003) Non-clinical safety study guideline. MHLW website: http://
wwwhourei.mhlw.go.jplhoureil (in Japanese)
OECD (2000) OECD Health data 2000 - a comparative analysis of 29 coun-
tries. OECD, Paris
Pharmaceutical Evaluation Study Group (2001a) Table of new chemical en-
tity drugs. In: Pharmaceutical manufacturing guidelines 2001 (in Japa-
nese). Jiho, Tokyo, pp 635-687
Parkinson C (1994) International survey on the timing of toxicity studies. I.
Relation to clinical trials. In: Parkinson C, McAuslane N, Lumley C,
Walker S (eds) The timing of toxicological studies to support clinical
trials. Proceedings of a CMR discussion meeting held at Nutfield Priory.
Kluwer Academic Publishers, DordrechtIBostonILondon, pp 27-39
Pharmaceutical Evaluation Study Group (2001 b) Handling of radiopharma-
ceuticals. In: Pharmaceutical manufacturing guidelines 2001 (in Japa-
nese). Jiho, Tokyo, pp 280-283
Sakai T, Takahashi K, Mitumori K, Yasuhara K, Kawashima K, Mayahara
H, Ohno Y (2000) Collaborative work to evaluate toxicity on male repro-
ductive organs by repeated dose studies in rats - overview of the studies.
J Toxicol Sci 25:1-21
Scales MDC, Mahoney K (1991) Animal toxicology studies on new medi-
cines and their relationship to clinical exposure: a review of international
recommendations. Adv Drug Reactions Toxicol Rev 10:155-168
Tajiri J (2003) http://www.j-tajiri.or.jp/
Takayama S, Akaike M, Kawashima K, Takahashi M, Kurokawa T (1995)
Studies on the optimal treatment period and parameters for detection of
male fertility disorder in rats. J Toxicol Sci 20: 173-193
Yahoo Japan News (2002) Yomiuri On Line (in Japanese), 25 Dec 2002.
http://www.yomiuri.co.jp/iryou/news_i120021225s011.htm
Ernst Schering Research Foundation Workshop
Editors: Gunter Stock

Monika Lessl
Vol. 1 (1991): Bioscience ~ Societly Workshop Report

Editors: D. J. Roy, B. E. Wynne, R. W. Old
Vol. 2 (1991): Round Table Discussion on Bioscience ~ Society
Editor: J. J. Cherfas
Vol. 3 (1991): Excitatory Amino Acids and Second Messenger Systems
Editors: V.1. Teichberg, L. Turski
Vol. 4 (1992): Spermatogenesis - Fertilization - Contraception
Editors: E. Nieschlag, U.-F. Habenicht
Vol. 5 (1992): Sex Steroids and the Cardiovascular System
Editors: P. Ramwell, G. Rubanyi, E. Schillinger
Vol. 6 (1993): Transgenic Animals as Model Systems for Human Diseases
Editors: E. F. Wagner, F. Theuring
Vol. 7 (1993): Basic Mechanisms Controlling Term and Preterm Birth
Editors: K. Chwalisz, R. E. Garfield
Vol. 8 (1994): Health Care 2010
Editors: C. Bezold, K. Knabner
Vol. 9 (1994): Sex Steroids and Bone
Editors: R. Ziegler, J. Pfeilschifter, M. Brautigam
Vol. 10 (1994): Nongenotoxic Carcinogenesis
Editors: A. Cockburn, L. Smith
Vol. 11 (1994): Cell Culture in Pharmaceutical Research
Editors: N. E. Fusenig, H. Graf
Vol. 12 (1994): Interactions Between Adjuvants, Agrochemical
and Target Organisms
Editors: P. J. Holloway, R. T. Rees, D. Stock
Vol. 13 (1994): Assessment of the Use of Single Cytochrome
P450 Enzymes in Drug Research
Editors: M. R. Waterman, M. Hildebrand
Vol. 14 (1995): Apoptosis in Hormone-Dependent Cancers
Editors: M. Tenniswood, H. Michna
Vol. 15 (1995): Computer Aided Drug Design in Industrial Research
Editors: E. C. Herrmann, R. Franke
Vol. 16 (1995): Organ-Selective Actions of Steroid Hormones
Editors: D. T. Baird, G. Schutz, R. Krattenmacher
Vol. 17 (1996): Alzheimer's Disease
Editors: J. D. Turner, K. Beyreuther, F. Theuring
Vol. 18 (1997): The Endometrium as a Target for Contraception
Editors: H.M. Beier, M.J.K. Harper, K. Chwalisz
Vol. 19 (1997): EGF Receptor in Tumor Growth and Progression
Editors: R. B. Lichtner, R. N. Harkins
Vol. 20 (1997): Cellular Therapy
Editors: H. Wekerle, H. Graf, J. D. Turner
Vol. 21 (1997): Nitric Oxide, Cytochromes P 450,
and Sexual Steroid Hormones
Editors: J. R. Lancaster, J. F. Parkinson
Vol. 22 (1997): Impact of Molecular Biology
and New Technical Developments in Diagnostic Imaging
Editors: W. Semmler, M. Schwaiger
Vol. 23 (1998): Excitatory Amino Acids
Editors: P. H. Seeburg, I. Bresink, L. Turski
Vol. 24 (1998): Molecular Basis of Sex Hormone Receptor Function
Editors: H. Gronemeyer, U. Fuhrmann, K. Parczyk
Vol. 25 (1998): Novel Approaches to Treatment of Osteoporosis
Editors: R. G. G. Russell, T. M. Skerry, U. Kollenkirchen
Vol. 26 (1998): Recent Trends in Molecular Recognition
Editors: F. Diederich, H. Kunzer
Vol. 27 (1998): Gene Therapy
Editors: R. E. Sobol, K. J. Scanlon, E. Nestaas, T. Strohmeyer
Vol. 28 (1999): Therapeutic Angiogenesis
Editors: J. A. Dormandy, W. P. Dole, G. M. Rubanyi
Vol. 29 (2000): Of Fish, Fly, Worm and Man
Editors: C. Nusslein-Volhard, J. Kratzschmar
Vol. 30 (2000): Therapeutic Vaccination Therapy
Editors: P. Walden, W. Sterry, H. Hennekes
Vol. 31 (2000): Advances in Eicosanoid Research
Editors: C. N. Serhan, H. D. Perez
Vol. 32 (2000): The Role of Natural Products in Drug Discovery
Editors: J. Mulzer, R. Bohlmann
Vol. 33 (2001): Stem Cells from Cord Blood, In Utero Stem Cell
Development, and Transplantation-Inclusive Gene Therapy
Editors:w. Holzgreve, M. Lessl
Vol. 34 (2001): Data Mining in Structural Biology
Editors: I. Schlichting, U. Egner
Vol. 35 (2002): Stem Cell Transplantation and Tissue Engineering
Editors: A. Haverich, H. Graf
Vol. 36 (2002): The Human Genome
Editors: A. Rosenthal, L. Vakalopoulou
Vol. 37 (2002): Pharmacokinetic Challenges in Drug Discovery
Editors: O. Pelkonen, A. Baumann, A. Reichel
Vol. 38 (2002): Bioinformatics and Genome Analysis
Editors: H.-W. Mewes, B. Weiss, H. Seidel
Vol. 39 (2002): Neuroinflammation - From Bench to Bedside
Editors: H. Kettenmann, G.A. Burton, U. Moenning
Vol. 40 (2002): Recent Advances in Glucocorticoid Receptor Action
Editors: A. Cato, H. Schaecke, K. Asadullah
Vol. 41 (2002): The Future of the Oocyte
Editors: J. Eppig, C. Hegele-Hartung
Vol. 42 (2003): Small Molecule-Protein Interaction
Editors: H. Waldmann, M. Koppitz
Vol. 43 (2003): Human Gene Therapy:
Present Opportunities and Future Trends
Editors: G. M. Rubanyi, S. Yla-Herttuala
Vol. 44 (2004): Leucocyte Trafficking:
The Role of Fucosyltransferases and Selectins
Editors: A. Hamann, K. Asadullah, A. Schottelius
Vol. 45 (2004): Chemokine Roles in Immunoregulation and Disease
Editors: P. M. Murphy, R. Horuk
Vol. 46 (2004): New Molecular Mechanisms of Estrogen Action
and Their Impact on Future Perspectives in Estrogen Therapy
Editors: K. S. Korach, A. Hillisch, K. H. Fritzemeier
Vol. 47 (2004): Neuroinflammation in Stroke
Editors: U. Dirnagl, B. Elger
Vol. 48 (2004): From Morphological Imaging to Molecular Targeting
Editors: M. Schwaiger, L. Dinkelborg, H. Schweinfurth
Supplement 1 (1994): Molecular and Cellular Endocrinology of the Testis
Editors: G. Verhoeven, U.-F. Habenicht
Supplement 2 (1997): Signal Transduction in Testicular Cells
Editors: V. Hansson, F. O. Levy, K. Tasken
Supplement 3 (1998): Testicular Function:
From Gene Expression to Genetic Manipulation
Editors: M. Stefanini, C. Boitani, M. Galdieri, R. Geremia,
F. Palombi
Supplement 4 (2000): Hormone Replacement Therapy
and Osteoporosis
Editors: J. Kato, H. Minaguchi, Y. Nishino
Supplement 5 (1999): Interferon: The Dawn of Recombinant
Protein Drugs
Editors: J. Lindenmann, W. D. Schleuning
Supplement 6 (2000): Testis, Epididymis and Technologies
in the Year 2000
Editors: B. Jegou, C. Pineau, J. Saez
Supplement 7 (2001): New Concepts in Pathology and
Treatment of Autoimmune Disorders
Editors: P. Pozzilli, C. Pozzilli, J.-F. Kapp
Supplement 8 (2001): New Pharmacological Approaches
to Reproductive Health and Healthy Ageing
Editors: W.-K. Raff, M. F. Fathalla, F. Saad
Supplement 9 (2002): Testicular Tangrams
Editors: F. F. G. Rommerts, K. J. Teerds
Supplement 10 (2002): Die Architektur des Lebens
Editors: G. Stock, M. Lessl

Ernst Schering Research Foundation Workshop 48 From Morphological Imaging To Molecular Targeting

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ernst Schering Research Foundation Workshop 48 From Morphological Imaging To Molecular Targeting

Uploaded by

Copyright:

Available Formats

Ernst Schering Research Foundation Workshop 48

From Morphological Imaging to Molecular Targeting

From Morphological Imaging

M. Schwaiger, L. Dinkelborg, H. Schweinfurth

Because of the current progress in molecular medicine (genomics,

more, such validation studies may allow for a better understanding

find that this is exactly what should apply to proof-of-concept stud-

Molecular Imaging: Dream or Reality?

2 Target Discovery and Validation

3 Noninvasive Imaging in Drug Discovery and Development

4 Internal Dose Assessment -

5 PET for Drug Development

6 Future Directions in Molecular Imaging

7 Regulatory Implications - The EU Perspective

8 Nonclinical Development of Radiopharmaceuticals:

9 The Japanese Perspective on Radiopharmaceuticals

Previous Volumes Published in This Series ............. 193

Stab in, M.G.

One has to try the impossible to achieve the possible.

1.1 Imaging Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Imaging has witnessed considerable advances during the last

computed tomography (SPECT) and positron emission tomography

1.1 Imaging Technologies

Many new and established imaging modalities compete for clinical

Signal Transduction Optical

tag et a1. 1998). The penetration of light in tissue is a complex diffu-

clinically and experimentally attractive information of structure, phy-

1.2 Imaging Targets

What are the currently employed strategies of molecular imaging

• Reporter Gene Imaging

ter substrate (FIAU), which is phosphorylated by tk and retained in

(NIS) proposed by Dai et al. represents a promising tool for radio-

1.3 Visualization of Target Proteins

Fig. 3. Visualization of somatostatin receptors overexpressed in neuroendo-

Fig. 4. Autoradiographic measurement of radioactivity distribution in a RIP-

al identification of areas with increased angiogenetic activity. Haubner

1.4 Assessment of Protein Function

A traditional application of tracer approaches is to measure protein

Fig.S. PET studies of knockout mice (GLUT4) showing delayed response

1.5 Biological Imaging as a Clinical Approach

Besides the established use of radiolabeled glucose (FDG) in assess-

strumentation. Future studies will have to define the specificity of

Fig. 6. PETtCT fusion of atherosclerotic plaque imaging with F-l8 deoxy-

In terms of biological actlvIty, FDG tissue uptake has recently

1.6 Monitoring of Therapy Response

The most widely appreciated role of functional imaging focuses on

tion of therapy. Weber et al. demonstrated that PET responders have

There is no question that molecular imaging represents a rapidly ex-

gral part in the drug evaluation process. This symposium is thought

Acknowledgements. The authors thank Gabriele Sonoda for her ex-

Bell J, Taylor-Robinson S (2000) Assessing gene expression in vivo: mag-

Dai G, Levy 0, Carrasco N (1996) Cloning and characterization of the thy-

Langen KJ, Pauleit D, Coenen HH (2002) 3-[(123)I]Iodo-alpha-methyl-L-tyr-

Spaepen K, Stroobants S, Dupont P, Vandenberghe P, Maertens J, Bormans

design, synthesis and preclinical evaluation of a novel F-18-labeled, car-

The pharmaceutical industry as we know it is still relatively young

In the 1980s, the ability to rapidly clone and manipulate genes

Another significant trend in the late twentieth century is the emer-

2.2 The Drug Discovery Process

Going from a research idea to a marketed drug takes from 10 to

2.3 Target Discovery

Most biological and medical research is still carried out in publicly

expression of a particular protein or its mRNA in healthy versus dis-